diff --git "a/related_34K/test_related_short_2404.17513v1.json" "b/related_34K/test_related_short_2404.17513v1.json" new file mode 100644--- /dev/null +++ "b/related_34K/test_related_short_2404.17513v1.json" @@ -0,0 +1,1426 @@ +[ + { + "url": "http://arxiv.org/abs/2404.17513v1", + "title": "A Comprehensive Evaluation on Event Reasoning of Large Language Models", + "abstract": "Event reasoning is a fundamental ability that underlies many applications. It\nrequires event schema knowledge to perform global reasoning and needs to deal\nwith the diversity of the inter-event relations and the reasoning paradigms.\nHow well LLMs accomplish event reasoning on various relations and reasoning\nparadigms remains unknown. To mitigate this disparity, we comprehensively\nevaluate the abilities of event reasoning of LLMs. We introduce a novel\nbenchmark EV2 for EValuation of EVent reasoning. EV2 consists of two levels of\nevaluation of schema and instance and is comprehensive in relations and\nreasoning paradigms. We conduct extensive experiments on EV2. We find that LLMs\nhave abilities to accomplish event reasoning but their performances are far\nfrom satisfactory. We also notice the imbalance of event reasoning abilities in\nLLMs. Besides, LLMs have event schema knowledge, however, they're not aligned\nwith humans on how to utilize the knowledge. Based on these findings, we\nintroduce two methods to guide the LLMs to utilize the event schema knowledge.\nBoth methods achieve improvements.", + "authors": "Zhengwei Tao, Zhi Jin, Yifan Zhang, Xiancai Chen, Xiaoying Bai, Yue Fang, Haiyan Zhao, Jia Li, Chongyang Tao", + "published": "2024-04-26", + "updated": "2024-04-26", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "label": "Original Paper", + "paper_cat": "LLM Fairness", + "gt": "Event Reasoning Du et al. (2022) aims to select the accurate cause or effect event from candidates. Zhou et al. (2019) serves as a dataset for event temporal reasoning. Current works present a scenario of incorporating counterfactual reasoning (Qin et al., 2019, 2020). In addition to singleevent relation reasoning, existing works also reason events according to diversified event relations (Poria et al., 2021; Han et al., 2021; Yang et al., 2022). Tao et al. (2023b) further unifies datasets of several event-inter relations to transfer event relational knowledge to unseen tasks. Predicting events necessitates the model to anticipate forthcoming occurrences grounded in the present context (Zhao, 2021). Mostafazadeh et al. (2016) employs a multiple-choice framework to predict future events by encompassing a diverse range of common-sense connections among events. Guan et al. (2019) establish a dataset oriented towards capturing event logic, enabling the generative prediction of future incidents. Evaluations for LLMs Evaluating the capacities of LLMs is the foundation of using and improving them. One group of research evaluates the general abilities of LLMs (Hendrycks et al., 2020; Zheng et al., 2023; Zhong et al., 2023; Bang et al., 2023) Besides, existing works evaluate LLMs in specific tasks (Bang et al., 2023; Bian et al., 2023; Gao et al., 2023; Wei et al., 2023; Li et al., 2024). Related to event reasoning, Yuan et al. (2023) evaluated the ability to solve event relation extraction. Tao et al. (2023a) present the Event Semantic Processing including the event understanding, reasoning, and prediction of event semantics. Chan et al. (2023) investigates relation reasoning between sentences. Compared with them, we are the first to introduce the evaluation for both schemaand instance-level event reasoning. Moreover, we comprehensively evaluate the performances of various relations and reasoning paradigms.", + "pre_questions": [], + "main_content": "Introduction Events are instances or occurrences that form the basic semantic building units encompassing the meanings of Activities, Accomplishments, Achievements, and States (Vendler, 1957). Event Reasoning is the ability to process and analyze events and their complex interconnections. Compared with other abilities, event reasoning is unique in some aspects. Firstly, it requires knowledge in the form of event schemas, capturing the progress of event evolution in scenarios, then performing global reasoning (Li et al., 2021a; Mao et al., 2021). **Corresponding authors. Jamie lay on a grassy hill, eyes closed, mind racing. Context CRR: Which is the relationship? Answer: Causes Jamie immersed in celestial physics at university. Jamie readied for an academic challenge. Before Before Dream Immersed in study, Jamie found tranquility in familiar stars. Calm Learn Become Ready CEC: Which happened after? Jamie sketched distant planets with a telescope. Explore Jamie chose theoretical astrophysics. Decide Jamie organized astronomy texts for efficient study. Organize Jamie focused on memorizing details. Memory Figure 1: An example of event reasoning. The red words are event schema knowledge. The sentences below are event instances. In event reasoning, there are various paradigms such as Contextual Event Classification (CEC) and Contextual Relation Reasoning (CRR), and diverse inter-event relations. As shown in Figure 1, each event instance is associated with an event type. All event types and their relations form the event schema knowledge which reflects the logic and mechanism of event evolution. Knowing \u201cMemory\u201d would often happen after \u201cLearn\u201d can help answer the reasoning question. Second, the inter-event relations and reasoning paradigms are various. Event reasoning incorporates reasoning events according to a certain relation (Du et al., 2022; Sap et al., 2019b) and reasoning inter-event relations (Ning et al., 2018; Caselli and Vossen, 2017). The queried relations are diversified such as causality (Roemmele et al., 2011), temporality (Zhou et al., 2019), and arXiv:2404.17513v1 [cs.CL] 26 Apr 2024 hierachy (Glava\u0161 et al., 2014). There are various paradigms such as reasoning the event or the interrelation. As a fundamental competency within LLMs, event reasoning supports a multitude of Natural Language Processing (NLP) tasks, including recommendation engines (Yang et al., 2020), interactive question-answer systems (Souza Costa et al., 2020), and AI Agents (Liu et al., 2023). Therefore, the enhancement of event reasoning abilities is essential for the advancement of LLMs. LLMs like LLAMA (Touvron et al., 2023) series and GPT series (Brown et al., 2020) have demonstrated exceptional accomplishments in various natural language reasoning (Bang et al., 2023; Xu et al., 2023b). Existing research has evaluated a broad spectrum of reasoning abilities of LLMs such as commonsence (Bian et al., 2023), sentence relations (Chan et al., 2023), and math (Arora et al., 2023). However, studies on the comprehensive evaluation of event reasoning of LLMs are scarce. Current works only focus on instance-level events, resulting in unclearness of how LLMs understand and utilize the event schema knowledge (Chan et al., 2023). Besides, they ignore the diversity of relations and paradigms (Yuan et al., 2023). These disparities hinge on the development of such crucial abilities of LLMs. In this paper, we comprehensively evaluate event reasoning in knowledge and abilities. Since there are existing datasets that are comprehensive in relations and paradigms, and can cover both levels of schema and instance, we introduce a novel benchmark EV2 for the EValuation of EVent reasoning. EV2 is featured in evaluating both aligned schema-level and instance-level. The schema-level evaluation investigates the event schema knowledge of LLMs while the instance-level testifies the event reasoning abilities. Besides, to evaluate event reasoning in various types of relation and reasoning paradigms, EV2 includes two event reasoning tasks, namely Contextual Event Classification (CEC) and Contextual Relation Reasoning (CRR) as shown in Figure 1. EV2 is constructed from both GPT generation and human annotation. Utilizing EV2, we comprehensively evaluate how well LLMs do event reasoning in terms of abilities and knowledge. Specifically, we mainly explore four research questions: 1) How proficient abilities of event reasoning do LLMs have? 2) To what extent do LLMs have the event schema knowledge? 3) Are LLMs aligned with humans in leveraging event schema knowledge? 4) Can LLMs perform better event reasoning with explicit guidance of leveraging event schema knowledge? We conduct extensive experiments on EV2 to answer these questions. The results provide insights into event reasoning that: 1) LLMs have the abilities of event reasoning, but are far from satisfactory and are imbalanced in different relations and reasoning paradigms. 2) LLMs have event schema knowledge. They can answer the schema-level questions with similar accuracy to the instance-level questions. However, the development of schema-level abilities falls behind those of instance-level. 3) LLMs are not aligned with humans in the aspect of leveraging event schema knowledge. 4) Based on the findings, we design two mentoring methods to guide the LLMs to utilize event schema knowledge. One is to directly add event schema knowledge to the prompt. The second is guiding in a chain-of-thought format. With the designed guidances for utilizing event schema knowledge, LLMs can perform better event reasoning. Especially with direct guidance, LLMs get significant improvements. We summarize our contributions as follows: \u2022 We evaluate event reasoning in both levels of schema and instance, and various relations and paradigms. \u2022 We construct a novel benchmark EV2 which features two levels of evaluation and comprehensive in relations and reasoning paradigms. We conduct extensive experiments to probe how LLMs perform event reasoning. \u2022 We conclude several insights. Based on our findings, we design mentoring methods to guide LLMs to utilize event schema knowledge which achieves improvements in event reasoning. 2 Problem Formulation Event reasoning is to anticipate the occurrences of certain relations or deduce interrelated correlations (Tao et al., 2023a). The relations encompass causality (Du et al., 2022), temporality (Zhou et al., 2019), and hierarchy (Glava\u0161 et al., 2014). Event reasoning requires comprehension of event schema knowledge. An event schema of a scenario is a schema-level graph Gs = (Vs, Es)1, where Vs is the set of event types and Es is the 1Superscript s represents schema level. set of relations between events. Each edge in Es is a relation triplet (Es i , R, Es j ) standing for that there is the relation R between Es i and Es j . With instantiation, we have the instance-level event graph Gi = (Vi, Ei)2. An instance event Ei has an event type Es but with detailed event arguments and context (Mitchell, 2005). The nodes and edges of these two graphs are corresponding, namely, each triplet in Gs has a corresponding triplet in Gi with the same inter-relation. In both levels, we consider totally six relation types, namely R \u2208{Causes, IsResult, Before, After, IsSubevent, HasSubevent}. EV2 consists of two event reasoning paradigms for both levels of schema and instance. The first is Contextual Event Classification (CEC) and the second is Contextualized Relation Reasoning (CRR). CEC Given graph G, either schemaor instancelevel, queried event E \u2208G, and target relation R, CEC requires the model to answer an event Ea: Ea = M(E, R, G, C). (1) M is the model, C is the candidate event set. CEC evaluates the model\u2019s comprehension of event semantics and structure. CRR Given graph G, either schemaor instancelevel, two queried events Ei, Ej \u2208G, CRR requires to determine the relation R between them: R = M(Ei, Ej, G). (2) CRR evaluates the understanding of event relations. In both schema and instance levels, EV2 has CEC and CRR tasks. Schema-level tasks require models to be rich in knowledge while tasks for instance need models to process detailed information. 3 Benchmark Construction To create the EV2 benchmark, we curate a comprehensive dataset through a three-stage process. Initially, the schema graph Gs is established. Then, GPT4 is employed to generate the instance graph Gi. Lastly, human annotators are tasked with creating questions from Gs and Gi. 3.1 Schema Graph Construction We leverage EECKG (Wang et al., 2022b) to ensure a diverse range of event types in our schema. EECKG combines rule-based reasoning 2Superscript i represents instance level. with crowdsourced insights, built on ConceptNet\u2019s structure. Nodes in EECKG represent verb phrases as events, and edges denote inter-event relations, focusing on Causes3, Before, and HasSubevent. Our objective mandates that the nodes within Gs should represent event types. Therefore, we filter EECKG nodes, removing concrete event instances. Preference is given to nodes with at most two words, as longer descriptions tend to include specific details. For events with fewer than two words, we use GPT4 to enhance our selection, ensuring the appropriate abstraction level for our schema graph with the following prompt: s ### Instructions: Determine which of the following candidate phrases are abstract and conceptual event types. We identify a subset of remaining events that are too generic. To refine the event selection, we also exclude the most frequent events from our subset to avoid generic events. We then dissect the interconnected EECKG into separate components, each representing a distinct scenario. To prevent semantic drift, we carefully control the size of each component. Starting from a node, we conduct a random walk until the number of nodes surpasses a set threshold, thus defining a component. This process is executed for all nodes to gather all components, as detailed in Algorithm 1. Post-extraction, we eliminate cycles to convert these structures into DAGs. EECKG only contains forward event evolution relations such as Causes. We further include components of backward relations. We generate a reversed version for each component by inverting edge directions and replacing relations with their opposites: IsResult, After, and IsSubevent. This creates the backward components. In preparation for constructing tasks for CEC and CRR, we label two events for each component. We sample three event pairs (Eh, Et) per component with a maximum inter-path length of four, utilizing their predecessors as background events. These pairs and background events form a schema graph. When the path length between Eh and Et is two, the direct relation serves as the queried relation; for longer paths, we deduce the relation using Table 1. We construct a schema graph, queried event pair, and their relation (Eh, Et, R, Gs). 3The direction is that the head event causes the tail event. Other relations are the same. Algorithm 1: Components Construction Input :EECKG G, N Output :A list of components O. 1 O = [ ] 2 Function RandomWalk(start, c): 3 l = RandomInt(N, N+2) 4 if l \u2264len(c) then 5 c.Append (start) 6 return c 7 n = Sample(start.Neighbors) 8 if n / \u2208c then 9 c \u2190RandomWalk(n, c \u222a{n}) 10 return c 11 return null 12 foreach node \u2208G do 13 component = RandomWalk (node, [ ]) 14 O.Append (component) 15 return O 3.2 Instance Graph Construction We next harvest instance graph Gi for each schema graph Gs. For each node Es \u2208Gs, we ask GPT4 to generate Ei using the following prompt: ### Instruction: Generate an instance event for each abstract event. The abstract event is the event type of the instance event. All the instance events form a coherent story which maintain the relations of each abstract event. The integrated story should have explicit roles, location and time. The whole story should be detailed, diverse in topic and scenarios, and rich in knowledge. We inherit the relations of Gs and obtain Gi. We naturally obtain the instances of Eh and Et. 3.3 Question Construction The last step is to construct questions of CEC and CRR in both schema and instance levels. For CEC, regarding schema and instance head events as the query and the tail as an answer, we ask GPT4 to generate 15 possible candidate instance events with their event types. We then recruit 8 well-educated human annotators. Their missions are: 1) Revise or discard Gs if not valid. Ensure the events are abstract, the relations are correct, there\u2019s no scenario shifting in Gs. 2) Revise or discard Gi if not valid. Ensure the events are concrete, the relations are correct, and the whole scenario of Gi is coherent and has no shifting. RULE INDUCTION (Before)+ Before (After)+ After (Before)\u22c6(Causes)+(Before)\u22c6 Before (After)\u22c6(IsResult)+(After)\u22c6 After (Before)\u22c6(HasSubevent)+( Before)\u22c6 Before (Causes)\u22c6(HasSubevent)+( Causes)\u22c6 Causes (After)\u22c6(IsSubevent)+( After)\u22c6 After (IsResult)\u22c6(IsSubevent)+( IsResult)\u22c6 IsResult Table 1: Relation induction rules. \u22c6denotes there exists zero or more. + means there is at least one. S-CEC I-CEC S-CRR S-CRR AVG N AVG E 492 558 767 835 3.62 2.78 Table 2: Statistic of EV2. S and I are schema and instance. AVG N and AVG E stand for the average number of nodes and edges per graph respectively. 3) Choose three proper negative candidate events with their event types. Ensure answering the question should rely on the context events. We use the schema part of annotation as the schemalevel questions and the instance part as instancelevel questions. Then we complete CEC. For CRR, we regard Es h and Es t as queried events and use the relation between them as the answer to form the schema-level question. For instance part, we adopt a similar way. Our CEC task is a 4-way multiple-choice task. The CRR is a 3-way multiple-choice task. In CRR, the choices for temporal, causal, and hierarchy relations are [Before, After, Vague], [Causes, IsResult, None], and [IsSubevent, HasSubevent, None] respectively. We show examples of both tasks in Figure 1. We report the number of each task and the average nodes and edges of EV2 in Figure 2. 3.4 Quality Inspection We recruit other human annotators to inspect the quality of EV2. We sample 100 data for all tasks. We ask them to give two scores for each sample: Correct: Rate 1 if correct, otherwise rate 0. Contextualized: Rate 1 if the answer relies on the context events, otherwise rate 0. Finally, we get 91% for Correct and 92% for Contextualized. Human examination testifies that EV2 is qualified. Besides, context events count. DATASET L C M-R M-P ALTLEX(Hidey, 2016) I \u2717 \u2717 \u2717 ASER(Zhang et al., 2020) S \u2717 \u2713 \u2717 ATOMIC(Sap et al., 2019a) S \u2717 \u2713 \u2717 COPA(Roemmele et al., 2011) I \u2717 \u2717 \u2717 CQA(Bondarenko et al., 2022) I \u2713 \u2713 \u2717 ECARE(Du et al., 2022) I \u2717 \u2717 \u2717 ESL(Caselli and Vossen, 2017) I \u2713 \u2717 \u2717 ESTER(Han et al., 2021) I \u2713 \u2713 \u2717 HIEVE(Glava\u0161 et al., 2014) I \u2713 \u2717 \u2717 KAIROS(Li et al., 2021a) S \u2713 \u2717 \u2717 LDC2020E25(Li et al., 2021a) S \u2713 \u2717 \u2717 MATRES(Ning et al., 2018) I \u2713 \u2717 \u2717 MAVEN-ERE(Wang et al., 2022a) I \u2713 \u2713 \u2717 MCNC(Granroth-Wilding, 2016) I \u2713 \u2717 \u2717 MCTACO(Zhou et al., 2019) I \u2713 \u2717 \u2717 RED(O\u2019Gorman et al., 2016) I \u2713 \u2713 \u2717 SCITE(Li et al., 2021b) I \u2713 \u2717 \u2717 SCT(Mostafazadeh et al., 2016) I \u2713 \u2717 \u2717 SocialIQA(Sap et al., 2019b) I \u2713 \u2713 \u2717 TB-Dense(Cassidy et al., 2014) I \u2713 \u2717 \u2717 TRACIE(Zhou et al., 2020) I \u2713 \u2717 \u2717 EV2 S I \u2713 \u2713 \u2713 Table 3: Comparison with existing event reasoning datasets. L stands for the included levels. C represents whether it\u2019s contextualized. M-R and M-P means if it has multi-relations and paradigms. S and I stand for schema and instance level. 3.5 Existing Dataset Comparison We compare our benchmark to existing related datasets. We show detailed comparison in Table 3. Our benchmark is the only one that is for contextualized event reasoning of various relations and paradigms on both schema and instance levels. 4 Experiments 4.1 Evaluated LLMs We evaluate 9 LLMs on event reasoning. For the open-source models, we evaluate their chat-version. For the closed-source models, we utilize their official APIs to conduct performance evaluations. Specifically, we employ the gpt-4-0125-preview version as the GPT4 and the gpt-3.5-turbo-1106 version as GPT3.5 in our experiments. For the opensource models, we include Qwen1.5-7B (Bai et al., 2023), Mistral-7B (Jiang et al., 2023), Baichuan-27B (Yang et al., 2023), Llama2-7B (Touvron et al., 2023), WizardLM-7B (Xu et al., 2023a), Vicuna7B (Chiang et al., 2023), and Alpaca-7B (Taori et al., 2023). Without loss of generosity, we use the model names to refer to the chat versions in the rest of our paper. For all evaluated LLMs, we use the same prompt. We show prompts in Figure 5-8 in the Appendix. 5 Results and Findings 5.1 How proficient abilities of event reasoning do LLMs have? In this part, we mainly probe the abilities of how existing LLMs complete the event reasoning of the instance level. LLMs have the abilities of event reasoning, but even the strongest GPT-4 is far from satisfactory. We evaluate CEC and CRR at the instance level. We show the results of different relations in Figure 2. For CEC, GPT4 performs the best. Models like Qwen1.5-7B, Mistral-7B, and GPT3.5 are in the second tier. Qwen1.5-7B and Mistral-7B are both better than GPT3.5. Qwen1.5-7B can even excel GPT4 in the temporal and causal relations. The other models such as WizardLM-7B almost fail, obtaining lower than 40% accuracy. For CRR, GPT4 excels all other models as well. However, unlike CEC, there is no obvious difference in the performance of other models for CRR. We show the average performance of instancelevel CEC and CRR in columns I-CEC and I-CRR in Table 4. We only show models that have basic abilities, namely CEC accuracy above 50.00 or CRR above 40.00 in Table 4, while other models may lack analytical significance. Overall, existing LLMs such as GPT4, and Qwen1.5-7B have CECtain event reasoning abilities. However, even the strongest GPT4 can only achieve 63.80 and 61.20 accuracy in each task showing there\u2019s much room for improvements of event reasoning. The abilities of LLMs to deal with different relations and reasoning paradigms are unbalanced. Comparing CEC to CRR, as relation-wise results shown in Figure 2 and average performances in columns I-CEC and I-CRR in Table 4, LLMs perform better for CEC than CRR. We compute the average scores of four listed models in Table 4. We find I-CEC is much higher than I-CRR, with 58.91 to 46.18. The results significantly suggest that CRR is harder than CEC. Existing pretraining and SFT datasets may be biased in paradigms. We then analyze performances on different relations. As shown in Figure 2, LLMs perform best in causality relation. Then, temporal, and hierarchy relations are tied. That further indicates the imbalance training of different relations. Methods and datasets of balanced abilities on relations are needed. Transferring abilities of different relations could also be feasible (Tao et al., 2023b). S-T S-C S-H I-T I-C I-H 20 30 40 50 60 70 CEC S-T S-C S-H I-T I-C I-H 20 30 40 50 60 70 CRR GPT4 GPT3.5 Qwen1.5-7B Mistral-7B Baichuan2-7B Llama2-7B WizardLM-7B Vicuna-7B Alpaca-7B Figure 2: Results of CEC and CRR. S and I stand for schemaand instance-level. Relation types of Causality, Temporality, and Hierarchy are denoted as C, T, and H. This is a crucial finding. Chan et al. (2023) conduct causal event classification such as ECARE (Du et al., 2022), and relation reasoning such as MATRES (Ning et al., 2018). They directly compare these two groups of results and conclude the gaps are merely from differences in relations. However, they ignore the difference in reasoning paradigms. Leveraging EV2, with disentangling relations and formulations, we investigate event reasoning with less bias. LLMs excel in forward CEC compared with backward. We calculate the average scores of forward relation (After, IsResult) and backward relation (Before, Causes). The results are shown in columns I-F, and I-B in Table 4. We find that the average of I-F is significantly better than I-B. It also suggests that the training dataset is unbalanced in relations. Less of the training data is used for backward relations, resulting in poorer performances on those. However, backward relations are important in abduction scenarios. Methods should be designed to enhance such abilities. CEC improves faster than CRR with model development. We investigate the improvement trends of CEC and CRR. In Figure 3. When models have poor event reasoning abilities, their performances lie around the balanced line showing no significant differences in tasks. With the development, the CEC improves much faster than CRR such models as GPT3.5, Mistral-7B, and Qwen1.57B. This investigation appeals to the need for training in comprehensive event reasoning abilities. Model S-F S-B I-F I-B S-CEC I-CEC S-CRR I-CRR GPT4 54.22 55.84 65.34 61.84 55.48 63.80 52.80 61.20 GPT3.5 47.88 50.94 55.68 49.12 49.79 50.18 45.37 39.52 Qwen 42.96 52.45 67.05 65.37 48.98 63.98 43.00 40.00 Mistral 44.37 52.08 67.61 53.71 48.98 57.71 46.00 44.00 AVG 47.35 52.82 63.92 57.51 50.8 58.91 46.79 46.18 Table 4: Average performances. AVG stands for the average scores of all models on that column. S and I stand for schemaand instance-level. F and B are forward and backward relations. 5.2 To what extent do LLMs have the event schema knowledge? In the previous section, we acknowledge that LLMs can complete event reasoning to some extent. However, whether they are endowed with event schema knowledge remains unknown. In this part, we mainly explore to what extent LLMs have the event schema knowledge, i.e. of the schema level. LLMs have event schema knowledge. We evaluate CEC and CRR on the schema level. The results are shown in Figure 2, and the average scores are reported in Table 4. We find LLMs already have event schema knowledge and can complete both CEC and CRR tasks at the schema level to some extent. However, in Table 4, we observe that S-CEC lags I-CEC, suggesting that LLMs are more adept at reasoning at the instance level. 20 30 40 50 60 70 CEC 30 35 40 45 50 55 60 65 CRR GPT4 GPT3.5 Mistral Qwen1.5 Baichuan2 Llama2 WizardLM Vicuna Alpaca Figure 3: Improvements trend on CEC and CRR. The dashline represents the balanced improvement with slope 3/4 considering the CEC is a 4-way multiplechoice task while CRR has three choices. The red line is the regression line of models except GPT4. Event schema knowledge increases falling behind reasoning at the instance level. We probe how event schema knowledge increases with the development of LLMs. We depict CEC performance comparisons of LLMs on instanceand schemalevel in Figure 4. When the models initially can reason about events, they also have event schema knowledge. At this time, models can perform comparatively or even better in schema-level event reasoning. With the development, models perform instance-level reasoning better than schema-level. It indicates that the accumulation of event schema knowledge falls behind the reasoning at the instance level. This finding demonstrates that enhancing event schema knowledge may further improve these abilities to obtain better general LLMs. 5.3 Are LLMs aligned with humans in the aspect of leveraging event schema knowledge In this section, we investigate how LLMs leverage event schema knowledge to complete event reasoning. We first provide the instance-level question for the models and then ask them to generate the required event schema knowledge to solve the task. Then we evaluate the accuracy of the generated event schema knowledge. Since we have the ground truth event schema knowledge for each question, the only challenge is to guide the LLMs to generate in a similar format for calculating accuracy. The instruction of our prompt first asks LLMs to generate the event types 20 30 40 50 60 70 I-CEC 20 25 30 35 40 45 50 55 60 S-CEC GPT4 GPT3.5 Qwen1.5 Mistral Llama2 WizardLM Vicuna Alpaca Figure 4: Comparisons between CEC performances on instanceand schema-level. The dashed line represents the balanced improvement with slope 1. The red line is the regression line of all models of each instance event in data. Based on the event types, it requires the LLMs to further generate relation triplets needed for the question. However, we find the LLMs would generate event types of different words but correct contents. To mitigate this problem, we prepare a list of candidate event types for each data to make it a classification setting. To keep the task difficult, we first conduct KMeans clustering on all event types in our dataset4. We obtain 1000 clusters. For each data, we assign 20 random candidates in total including the correct ones. The negative event types are chosen from different clusters. After the generation, we calculate the accuracy of event types and F1-scores of relation triplets respectively comparing with the human-labeled event schema. We regard a correct triplet if all the head and tail event types and the inter-relation align with the human labels. We show detailed examples in Figures 9-10 in Appendix. The results are in Table 5. We find only GPT4 can generate correct event types while other models all fail. For relation triplet generation, even GPT4 can not output proper event schemas5. It significantly suggests that LLMs may not leverage event schema knowledge as humans when solving event reasoning tasks. Alignment of using such knowledge could further improve the performances. 4We use all-mpnet-base-v2 for encoding. 5GPT4 excels other may be attributed to 1) its better alignment. 2) The dataset is originally generated by GPT4. CEC CRR ET REL ET REL GPT4 70.71 37.30 70.66 49.41 GPT3.5 13.43 15.78 18.55 21.14 Mistral-7B 11.15 9.00 11.88 15.15 Table 5: Event schema knowledge Alignment. ET is the event type accuracy. REL is relation triplet F1-score. CEC CRR W.O.S W.T.S W.O.S W.T.S GPT4 63.80 69.89 (6.09\u2191) 61.2 63.11 (1.91\u2191) GPT3.5 50.18 60.92 (10.74\u2191) 39.52 45.99 (6.47\u2191) Mistral-7B 57.71 63.26 (5.55\u2191) 44.00 47.07 (3.07\u2191) Llama2-7B 30.29 38.17 (7.88\u2191) 34.00 43.35 (9.35\u2191) WizardLM 33.69 29.93 (3.76\u2193) 37.00 44.91 (7.91\u2191) Vicuna-7B 31.18 34.41 (3.23\u2191) 42.00 42.40 (0.40\u2191) Table 6: Direct guidance with schema knowledge. W.T.S and W.O.S stands for with and without event knowledge guidance. We also report the difference between them. 5.4 Can LLMs perform better event reasoning with explicit guidance of leveraging event schema knowledge? In the previous section, we find LLMs may not leverage event schema knowledge as human does. It raises an interesting question how well LLMs perform if we guide them to explicitly use such knowledge? In this section, we probe this question. We design two guiding methods: Direct: Directly add the event type of each instance event into the prompt. CoT: Guide the LLMs in a CoT-style to 1) generate the event types of each instance event. 2) reason with the event types. This is a more practical method since we would not know the event types in advance in real scenarios. We show the performances of direct guidance in Table 6. We find incorporating event schema knowledge significantly improves event reasoning. It shows great potential to solve event reasoning with the fusion of event schema knowledge. We report the results of the CoT guidance in Table 7. We only report results of GPT-4 since we find other models are unable to follow this instruction. We find in CEC, CoT guidance can improve performance. However, the improvement of CoT lags those of Direct, indicating great space for better methods. Developing advanced guidance for all LLMs remains a challenging research problem. We CEC W.O.S W.T.S GPT4 63.80 67.92 (4.12\u2191) Table 7: CoT guidance with schema knowledge. show the example of prompt and GPT4 generation in Figures 11 and 12 in Appendix. In this paper, we evaluate the event reasoning of LLMs. We introduce a novel benchmark EV2 which features both levels of schema and instance. It evaluates event schema knowledge and reasoning abilities. Besides, EV2 can be used to comprehensively evaluate the event reasoning in various relations and reasoning paradigms. We conduct extensive experiments on EV2. We obtain many insights such as: 1) LLMs have the abilities of event reasoning, but are far from satisfactory and are unbalanced in different relations and reasoning paradigms. 2) LLMs have event schema knowledge. However, with the development of LLMs, this knowledge increases slowly compared with the increase of abilities of event instance reasoning. 3) LLMs are not aligned with human to leaverage event schema knowledge in event reasoning. 4) Based on the findings, we design two methods, namely Direct and CoT, to guide the LLMs to utilize event schema knowledge. With our designed guidances for utilizing event schema knowledge, LLMs can perform better event reasoning. Limitations We guide the LLMs to utilize the event schema knowledge in two ways. The Direct effects most. However, the more practical way CoT falls behind Direct indicating the potential of a better method of guidance. We leave it to future work." + }, + { + "url": "http://arxiv.org/abs/2205.05849v1", + "title": "e-CARE: a New Dataset for Exploring Explainable Causal Reasoning", + "abstract": "Understanding causality has vital importance for various Natural Language\nProcessing (NLP) applications. Beyond the labeled instances, conceptual\nexplanations of the causality can provide deep understanding of the causal\nfacts to facilitate the causal reasoning process. However, such explanation\ninformation still remains absent in existing causal reasoning resources. In\nthis paper, we fill this gap by presenting a human-annotated explainable CAusal\nREasoning dataset (e-CARE), which contains over 21K causal reasoning questions,\ntogether with natural language formed explanations of the causal questions.\nExperimental results show that generating valid explanations for causal facts\nstill remains especially challenging for the state-of-the-art models, and the\nexplanation information can be helpful for promoting the accuracy and stability\nof causal reasoning models.", + "authors": "Li Du, Xiao Ding, Kai Xiong, Ting Liu, Bing Qin", + "published": "2022-05-12", + "updated": "2022-05-12", + "primary_cat": "cs.AI", + "cats": [ + "cs.AI", + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2012.11820v4", + "title": "Recognizing Emotion Cause in Conversations", + "abstract": "We address the problem of recognizing emotion cause in conversations, define\ntwo novel sub-tasks of this problem, and provide a corresponding dialogue-level\ndataset, along with strong Transformer-based baselines. The dataset is\navailable at https://github.com/declare-lab/RECCON.\n Introduction: Recognizing the cause behind emotions in text is a fundamental\nyet under-explored area of research in NLP. Advances in this area hold the\npotential to improve interpretability and performance in affect-based models.\nIdentifying emotion causes at the utterance level in conversations is\nparticularly challenging due to the intermingling dynamics among the\ninterlocutors.\n Method: We introduce the task of Recognizing Emotion Cause in CONversations\nwith an accompanying dataset named RECCON, containing over 1,000 dialogues and\n10,000 utterance cause-effect pairs. Furthermore, we define different cause\ntypes based on the source of the causes, and establish strong Transformer-based\nbaselines to address two different sub-tasks on this dataset: causal span\nextraction and causal emotion entailment.\n Result: Our Transformer-based baselines, which leverage contextual\npre-trained embeddings, such as RoBERTa, outperform the state-of-the-art\nemotion cause extraction approaches\n Conclusion: We introduce a new task highly relevant for (explainable)\nemotion-aware artificial intelligence: recognizing emotion cause in\nconversations, provide a new highly challenging publicly available\ndialogue-level dataset for this task, and give strong baseline results on this\ndataset.", + "authors": "Soujanya Poria, Navonil Majumder, Devamanyu Hazarika, Deepanway Ghosal, Rishabh Bhardwaj, Samson Yu Bai Jian, Pengfei Hong, Romila Ghosh, Abhinaba Roy, Niyati Chhaya, Alexander Gelbukh, Rada Mihalcea", + "published": "2020-12-22", + "updated": "2021-07-28", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2303.16421v3", + "title": "ChatGPT is a Knowledgeable but Inexperienced Solver: An Investigation of Commonsense Problem in Large Language Models", + "abstract": "Large language models (LLMs) have made significant progress in NLP. However,\ntheir ability to memorize, represent, and leverage commonsense knowledge has\nbeen a well-known pain point. In this paper, we specifically focus on ChatGPT,\na widely used and easily accessible LLM, and ask the following questions: (1)\nCan ChatGPT effectively answer commonsense questions? (2) Is ChatGPT aware of\nthe underlying commonsense knowledge for answering a specific question? (3) Is\nChatGPT knowledgeable in commonsense? (4) Can ChatGPT effectively leverage\ncommonsense for answering questions? We conduct a series of experiments on 11\ndatasets to evaluate ChatGPT's commonsense abilities, including answering\ncommonsense questions, identifying necessary knowledge, generating knowledge\ndescriptions, and using knowledge descriptions to answer questions again.\nExperimental results show that: (1) ChatGPT can achieve good QA accuracies in\ncommonsense tasks, while still struggling with certain domains of datasets. (2)\nChatGPT is knowledgeable, and can accurately generate most of the commonsense\nknowledge using knowledge prompts. (3) Despite its knowledge, ChatGPT is an\ninexperienced commonsense problem solver, which cannot precisely identify the\nneeded commonsense for answering a specific question. These findings raise the\nneed to explore improved mechanisms for effectively incorporating commonsense\ninto LLMs like ChatGPT, such as better instruction following and commonsense\nguidance.", + "authors": "Ning Bian, Xianpei Han, Le Sun, Hongyu Lin, Yaojie Lu, Ben He, Shanshan Jiang, Bin Dong", + "published": "2023-03-29", + "updated": "2024-04-19", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2304.06364v2", + "title": "AGIEval: A Human-Centric Benchmark for Evaluating Foundation Models", + "abstract": "Evaluating the general abilities of foundation models to tackle human-level\ntasks is a vital aspect of their development and application in the pursuit of\nArtificial General Intelligence (AGI). Traditional benchmarks, which rely on\nartificial datasets, may not accurately represent human-level capabilities. In\nthis paper, we introduce AGIEval, a novel benchmark specifically designed to\nassess foundation model in the context of human-centric standardized exams,\nsuch as college entrance exams, law school admission tests, math competitions,\nand lawyer qualification tests. We evaluate several state-of-the-art foundation\nmodels, including GPT-4, ChatGPT, and Text-Davinci-003, using this benchmark.\nImpressively, GPT-4 surpasses average human performance on SAT, LSAT, and math\ncompetitions, attaining a 95% accuracy rate on the SAT Math test and a 92.5%\naccuracy on the English test of the Chinese national college entrance exam.\nThis demonstrates the extraordinary performance of contemporary foundation\nmodels. In contrast, we also find that GPT-4 is less proficient in tasks that\nrequire complex reasoning or specific domain knowledge. Our comprehensive\nanalyses of model capabilities (understanding, knowledge, reasoning, and\ncalculation) reveal these models' strengths and limitations, providing valuable\ninsights into future directions for enhancing their general capabilities. By\nconcentrating on tasks pertinent to human cognition and decision-making, our\nbenchmark delivers a more meaningful and robust evaluation of foundation\nmodels' performance in real-world scenarios. The data, code, and all model\noutputs are released in https://github.com/ruixiangcui/AGIEval.", + "authors": "Wanjun Zhong, Ruixiang Cui, Yiduo Guo, Yaobo Liang, Shuai Lu, Yanlin Wang, Amin Saied, Weizhu Chen, Nan Duan", + "published": "2023-04-13", + "updated": "2023-09-18", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2401.07103v1", + "title": "Leveraging Large Language Models for NLG Evaluation: A Survey", + "abstract": "In the rapidly evolving domain of Natural Language Generation (NLG)\nevaluation, introducing Large Language Models (LLMs) has opened new avenues for\nassessing generated content quality, e.g., coherence, creativity, and context\nrelevance. This survey aims to provide a thorough overview of leveraging LLMs\nfor NLG evaluation, a burgeoning area that lacks a systematic analysis. We\npropose a coherent taxonomy for organizing existing LLM-based evaluation\nmetrics, offering a structured framework to understand and compare these\nmethods. Our detailed exploration includes critically assessing various\nLLM-based methodologies, as well as comparing their strengths and limitations\nin evaluating NLG outputs. By discussing unresolved challenges, including bias,\nrobustness, domain-specificity, and unified evaluation, this survey seeks to\noffer insights to researchers and advocate for fairer and more advanced NLG\nevaluation techniques.", + "authors": "Zhen Li, Xiaohan Xu, Tao Shen, Can Xu, Jia-Chen Gu, Chongyang Tao", + "published": "2024-01-13", + "updated": "2024-01-13", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2305.15268v1", + "title": "EvEval: A Comprehensive Evaluation of Event Semantics for Large Language Models", + "abstract": "Events serve as fundamental units of occurrence within various contexts. The\nprocessing of event semantics in textual information forms the basis of\nnumerous natural language processing (NLP) applications. Recent studies have\nbegun leveraging large language models (LLMs) to address event semantic\nprocessing. However, the extent that LLMs can effectively tackle these\nchallenges remains uncertain. Furthermore, the lack of a comprehensive\nevaluation framework for event semantic processing poses a significant\nchallenge in evaluating these capabilities. In this paper, we propose an\noverarching framework for event semantic processing, encompassing\nunderstanding, reasoning, and prediction, along with their fine-grained\naspects. To comprehensively evaluate the event semantic processing abilities of\nmodels, we introduce a novel benchmark called EVEVAL. We collect 8 datasets\nthat cover all aspects of event semantic processing. Extensive experiments are\nconducted on EVEVAL, leading to several noteworthy findings based on the\nobtained results.", + "authors": "Zhengwei Tao, Zhi Jin, Xiaoying Bai, Haiyan Zhao, Yanlin Feng, Jia Li, Wenpeng Hu", + "published": "2023-05-24", + "updated": "2023-05-24", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2303.03836v2", + "title": "Exploring the Feasibility of ChatGPT for Event Extraction", + "abstract": "Event extraction is a fundamental task in natural language processing that\ninvolves identifying and extracting information about events mentioned in text.\nHowever, it is a challenging task due to the lack of annotated data, which is\nexpensive and time-consuming to obtain. The emergence of large language models\n(LLMs) such as ChatGPT provides an opportunity to solve language tasks with\nsimple prompts without the need for task-specific datasets and fine-tuning.\nWhile ChatGPT has demonstrated impressive results in tasks like machine\ntranslation, text summarization, and question answering, it presents challenges\nwhen used for complex tasks like event extraction. Unlike other tasks, event\nextraction requires the model to be provided with a complex set of instructions\ndefining all event types and their schemas. To explore the feasibility of\nChatGPT for event extraction and the challenges it poses, we conducted a series\nof experiments. Our results show that ChatGPT has, on average, only 51.04% of\nthe performance of a task-specific model such as EEQA in long-tail and complex\nscenarios. Our usability testing experiments indicate that ChatGPT is not\nrobust enough, and continuous refinement of the prompt does not lead to stable\nperformance improvements, which can result in a poor user experience. Besides,\nChatGPT is highly sensitive to different prompt styles.", + "authors": "Jun Gao, Huan Zhao, Changlong Yu, Ruifeng Xu", + "published": "2023-03-07", + "updated": "2023-03-09", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2302.10205v1", + "title": "Zero-Shot Information Extraction via Chatting with ChatGPT", + "abstract": "Zero-shot information extraction (IE) aims to build IE systems from the\nunannotated text. It is challenging due to involving little human intervention.\nChallenging but worthwhile, zero-shot IE reduces the time and effort that data\nlabeling takes. Recent efforts on large language models (LLMs, e.g., GPT-3,\nChatGPT) show promising performance on zero-shot settings, thus inspiring us to\nexplore prompt-based methods. In this work, we ask whether strong IE models can\nbe constructed by directly prompting LLMs. Specifically, we transform the\nzero-shot IE task into a multi-turn question-answering problem with a two-stage\nframework (ChatIE). With the power of ChatGPT, we extensively evaluate our\nframework on three IE tasks: entity-relation triple extract, named entity\nrecognition, and event extraction. Empirical results on six datasets across two\nlanguages show that ChatIE achieves impressive performance and even surpasses\nsome full-shot models on several datasets (e.g., NYT11-HRL). We believe that\nour work could shed light on building IE models with limited resources.", + "authors": "Xiang Wei, Xingyu Cui, Ning Cheng, Xiaobin Wang, Xin Zhang, Shen Huang, Pengjun Xie, Jinan Xu, Yufeng Chen, Meishan Zhang, Yong Jiang, Wenjuan Han", + "published": "2023-02-20", + "updated": "2023-02-20", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2302.04023v4", + "title": "A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivity", + "abstract": "This paper proposes a framework for quantitatively evaluating interactive\nLLMs such as ChatGPT using publicly available data sets. We carry out an\nextensive technical evaluation of ChatGPT using 23 data sets covering 8\ndifferent common NLP application tasks. We evaluate the multitask, multilingual\nand multi-modal aspects of ChatGPT based on these data sets and a newly\ndesigned multimodal dataset. We find that ChatGPT outperforms LLMs with\nzero-shot learning on most tasks and even outperforms fine-tuned models on some\ntasks. We find that it is better at understanding non-Latin script languages\nthan generating them. It is able to generate multimodal content from textual\nprompts, via an intermediate code generation step. Moreover, we find that\nChatGPT is 63.41% accurate on average in 10 different reasoning categories\nunder logical reasoning, non-textual reasoning, and commonsense reasoning,\nhence making it an unreliable reasoner. It is, for example, better at deductive\nthan inductive reasoning. ChatGPT suffers from hallucination problems like\nother LLMs and it generates more extrinsic hallucinations from its parametric\nmemory as it does not have access to an external knowledge base. Finally, the\ninteractive feature of ChatGPT enables human collaboration with the underlying\nLLM to improve its performance, i.e, 8% ROUGE-1 on summarization and 2% ChrF++\non machine translation, in a multi-turn \"prompt engineering\" fashion. We also\nrelease codebase for evaluation set extraction.", + "authors": "Yejin Bang, Samuel Cahyawijaya, Nayeon Lee, Wenliang Dai, Dan Su, Bryan Wilie, Holy Lovenia, Ziwei Ji, Tiezheng Yu, Willy Chung, Quyet V. Do, Yan Xu, Pascale Fung", + "published": "2023-02-08", + "updated": "2023-11-28", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/1909.04076v2", + "title": "Counterfactual Story Reasoning and Generation", + "abstract": "Counterfactual reasoning requires predicting how alternative events, contrary\nto what actually happened, might have resulted in different outcomes. Despite\nbeing considered a necessary component of AI-complete systems, few resources\nhave been developed for evaluating counterfactual reasoning in narratives.\n In this paper, we propose Counterfactual Story Rewriting: given an original\nstory and an intervening counterfactual event, the task is to minimally revise\nthe story to make it compatible with the given counterfactual event. Solving\nthis task will require deep understanding of causal narrative chains and\ncounterfactual invariance, and integration of such story reasoning capabilities\ninto conditional language generation models.\n We present TimeTravel, a new dataset of 29,849 counterfactual rewritings,\neach with the original story, a counterfactual event, and human-generated\nrevision of the original story compatible with the counterfactual event.\nAdditionally, we include 80,115 counterfactual \"branches\" without a rewritten\nstoryline to support future work on semi- or un-supervised approaches to\ncounterfactual story rewriting.\n Finally, we evaluate the counterfactual rewriting capacities of several\ncompetitive baselines based on pretrained language models, and assess whether\ncommon overlap and model-based automatic metrics for text generation correlate\nwell with human scores for counterfactual rewriting.", + "authors": "Lianhui Qin, Antoine Bosselut, Ari Holtzman, Chandra Bhagavatula, Elizabeth Clark, Yejin Choi", + "published": "2019-09-09", + "updated": "2019-09-12", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2306.05685v4", + "title": "Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena", + "abstract": "Evaluating large language model (LLM) based chat assistants is challenging\ndue to their broad capabilities and the inadequacy of existing benchmarks in\nmeasuring human preferences. To address this, we explore using strong LLMs as\njudges to evaluate these models on more open-ended questions. We examine the\nusage and limitations of LLM-as-a-judge, including position, verbosity, and\nself-enhancement biases, as well as limited reasoning ability, and propose\nsolutions to mitigate some of them. We then verify the agreement between LLM\njudges and human preferences by introducing two benchmarks: MT-bench, a\nmulti-turn question set; and Chatbot Arena, a crowdsourced battle platform. Our\nresults reveal that strong LLM judges like GPT-4 can match both controlled and\ncrowdsourced human preferences well, achieving over 80% agreement, the same\nlevel of agreement between humans. Hence, LLM-as-a-judge is a scalable and\nexplainable way to approximate human preferences, which are otherwise very\nexpensive to obtain. Additionally, we show our benchmark and traditional\nbenchmarks complement each other by evaluating several variants of LLaMA and\nVicuna. The MT-bench questions, 3K expert votes, and 30K conversations with\nhuman preferences are publicly available at\nhttps://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge.", + "authors": "Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric P. Xing, Hao Zhang, Joseph E. Gonzalez, Ion Stoica", + "published": "2023-06-09", + "updated": "2023-12-24", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/1909.03065v1", + "title": "\"Going on a vacation\" takes longer than \"Going for a walk\": A Study of Temporal Commonsense Understanding", + "abstract": "Understanding time is crucial for understanding events expressed in natural\nlanguage. Because people rarely say the obvious, it is often necessary to have\ncommonsense knowledge about various temporal aspects of events, such as\nduration, frequency, and temporal order. However, this important problem has so\nfar received limited attention. This paper systematically studies this temporal\ncommonsense problem. Specifically, we define five classes of temporal\ncommonsense, and use crowdsourcing to develop a new dataset, MCTACO, that\nserves as a test set for this task. We find that the best current methods used\non MCTACO are still far behind human performance, by about 20%, and discuss\nseveral directions for improvement. We hope that the new dataset and our study\nhere can foster more future research on this topic.", + "authors": "Ben Zhou, Daniel Khashabi, Qiang Ning, Dan Roth", + "published": "2019-09-06", + "updated": "2019-09-06", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2304.14827v3", + "title": "ChatGPT Evaluation on Sentence Level Relations: A Focus on Temporal, Causal, and Discourse Relations", + "abstract": "This paper aims to quantitatively evaluate the performance of ChatGPT, an\ninteractive large language model, on inter-sentential relations such as\ntemporal relations, causal relations, and discourse relations. Given ChatGPT's\npromising performance across various tasks, we proceed to carry out thorough\nevaluations on the whole test sets of 11 datasets, including temporal and\ncausal relations, PDTB2.0-based, and dialogue-based discourse relations. To\nensure the reliability of our findings, we employ three tailored prompt\ntemplates for each task, including the zero-shot prompt template, zero-shot\nprompt engineering (PE) template, and in-context learning (ICL) prompt\ntemplate, to establish the initial baseline scores for all popular\nsentence-pair relation classification tasks for the first time. Through our\nstudy, we discover that ChatGPT exhibits exceptional proficiency in detecting\nand reasoning about causal relations, albeit it may not possess the same level\nof expertise in identifying the temporal order between two events. While it is\ncapable of identifying the majority of discourse relations with existing\nexplicit discourse connectives, the implicit discourse relation remains a\nformidable challenge. Concurrently, ChatGPT demonstrates subpar performance in\nthe dialogue discourse parsing task that requires structural understanding in a\ndialogue before being aware of the discourse relation.", + "authors": "Chunkit Chan, Jiayang Cheng, Weiqi Wang, Yuxin Jiang, Tianqing Fang, Xin Liu, Yangqiu Song", + "published": "2023-04-28", + "updated": "2024-01-26", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2204.07408v1", + "title": "Towards Fine-grained Causal Reasoning and QA", + "abstract": "Understanding causality is key to the success of NLP applications, especially\nin high-stakes domains. Causality comes in various perspectives such as enable\nand prevent that, despite their importance, have been largely ignored in the\nliterature. This paper introduces a novel fine-grained causal reasoning dataset\nand presents a series of novel predictive tasks in NLP, such as causality\ndetection, event causality extraction, and Causal QA. Our dataset contains\nhuman annotations of 25K cause-effect event pairs and 24K question-answering\npairs within multi-sentence samples, where each can have multiple causal\nrelationships. Through extensive experiments and analysis, we show that the\ncomplex relations in our dataset bring unique challenges to state-of-the-art\nmethods across all three tasks and highlight potential research opportunities,\nespecially in developing \"causal-thinking\" methods.", + "authors": "Linyi Yang, Zhen Wang, Yuxiang Wu, Jie Yang, Yue Zhang", + "published": "2022-04-15", + "updated": "2022-04-15", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.LO" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2309.10305v2", + "title": "Baichuan 2: Open Large-scale Language Models", + "abstract": "Large language models (LLMs) have demonstrated remarkable performance on a\nvariety of natural language tasks based on just a few examples of natural\nlanguage instructions, reducing the need for extensive feature engineering.\nHowever, most powerful LLMs are closed-source or limited in their capability\nfor languages other than English. In this technical report, we present Baichuan\n2, a series of large-scale multilingual language models containing 7 billion\nand 13 billion parameters, trained from scratch, on 2.6 trillion tokens.\nBaichuan 2 matches or outperforms other open-source models of similar size on\npublic benchmarks like MMLU, CMMLU, GSM8K, and HumanEval. Furthermore, Baichuan\n2 excels in vertical domains such as medicine and law. We will release all\npre-training model checkpoints to benefit the research community in better\nunderstanding the training dynamics of Baichuan 2.", + "authors": "Aiyuan Yang, Bin Xiao, Bingning Wang, Borong Zhang, Ce Bian, Chao Yin, Chenxu Lv, Da Pan, Dian Wang, Dong Yan, Fan Yang, Fei Deng, Feng Wang, Feng Liu, Guangwei Ai, Guosheng Dong, Haizhou Zhao, Hang Xu, Haoze Sun, Hongda Zhang, Hui Liu, Jiaming Ji, Jian Xie, JunTao Dai, Kun Fang, Lei Su, Liang Song, Lifeng Liu, Liyun Ru, Luyao Ma, Mang Wang, Mickel Liu, MingAn Lin, Nuolan Nie, Peidong Guo, Ruiyang Sun, Tao Zhang, Tianpeng Li, Tianyu Li, Wei Cheng, Weipeng Chen, Xiangrong Zeng, Xiaochuan Wang, Xiaoxi Chen, Xin Men, Xin Yu, Xuehai Pan, Yanjun Shen, Yiding Wang, Yiyu Li, Youxin Jiang, Yuchen Gao, Yupeng Zhang, Zenan Zhou, Zhiying Wu", + "published": "2023-09-19", + "updated": "2023-09-20", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2304.05454v1", + "title": "Zero-shot Temporal Relation Extraction with ChatGPT", + "abstract": "The goal of temporal relation extraction is to infer the temporal relation\nbetween two events in the document. Supervised models are dominant in this\ntask. In this work, we investigate ChatGPT's ability on zero-shot temporal\nrelation extraction. We designed three different prompt techniques to break\ndown the task and evaluate ChatGPT. Our experiments show that ChatGPT's\nperformance has a large gap with that of supervised methods and can heavily\nrely on the design of prompts. We further demonstrate that ChatGPT can infer\nmore small relation classes correctly than supervised methods. The current\nshortcomings of ChatGPT on temporal relation extraction are also discussed in\nthis paper. We found that ChatGPT cannot keep consistency during temporal\ninference and it fails in actively long-dependency temporal inference.", + "authors": "Chenhan Yuan, Qianqian Xie, Sophia Ananiadou", + "published": "2023-04-11", + "updated": "2023-04-11", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2009.03300v3", + "title": "Measuring Massive Multitask Language Understanding", + "abstract": "We propose a new test to measure a text model's multitask accuracy. The test\ncovers 57 tasks including elementary mathematics, US history, computer science,\nlaw, and more. To attain high accuracy on this test, models must possess\nextensive world knowledge and problem solving ability. We find that while most\nrecent models have near random-chance accuracy, the very largest GPT-3 model\nimproves over random chance by almost 20 percentage points on average. However,\non every one of the 57 tasks, the best models still need substantial\nimprovements before they can reach expert-level accuracy. Models also have\nlopsided performance and frequently do not know when they are wrong. Worse,\nthey still have near-random accuracy on some socially important subjects such\nas morality and law. By comprehensively evaluating the breadth and depth of a\nmodel's academic and professional understanding, our test can be used to\nanalyze models across many tasks and to identify important shortcomings.", + "authors": "Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, Jacob Steinhardt", + "published": "2020-09-07", + "updated": "2021-01-12", + "primary_cat": "cs.CY", + "cats": [ + "cs.CY", + "cs.AI", + "cs.CL", + "cs.LG" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/1808.10113v3", + "title": "Story Ending Generation with Incremental Encoding and Commonsense Knowledge", + "abstract": "Generating a reasonable ending for a given story context, i.e., story ending\ngeneration, is a strong indication of story comprehension. This task requires\nnot only to understand the context clues which play an important role in\nplanning the plot but also to handle implicit knowledge to make a reasonable,\ncoherent story.\n In this paper, we devise a novel model for story ending generation. The model\nadopts an incremental encoding scheme to represent context clues which are\nspanning in the story context. In addition, commonsense knowledge is applied\nthrough multi-source attention to facilitate story comprehension, and thus to\nhelp generate coherent and reasonable endings. Through building context clues\nand using implicit knowledge, the model is able to produce reasonable story\nendings. context clues implied in the post and make the inference based on it.\n Automatic and manual evaluation shows that our model can generate more\nreasonable story endings than state-of-the-art baselines.", + "authors": "Jian Guan, Yansen Wang, Minlie Huang", + "published": "2018-08-30", + "updated": "2018-12-02", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "label": "Related Work" + }, + { + "url": "http://arxiv.org/abs/2311.00306v1", + "title": "Probing Explicit and Implicit Gender Bias through LLM Conditional Text Generation", + "abstract": "Large Language Models (LLMs) can generate biased and toxic responses. Yet\nmost prior work on LLM gender bias evaluation requires predefined\ngender-related phrases or gender stereotypes, which are challenging to be\ncomprehensively collected and are limited to explicit bias evaluation. In\naddition, we believe that instances devoid of gender-related language or\nexplicit stereotypes in inputs can still induce gender bias in LLMs. Thus, in\nthis work, we propose a conditional text generation mechanism without the need\nfor predefined gender phrases and stereotypes. This approach employs three\ntypes of inputs generated through three distinct strategies to probe LLMs,\naiming to show evidence of explicit and implicit gender biases in LLMs. We also\nutilize explicit and implicit evaluation metrics to evaluate gender bias in\nLLMs under different strategies. Our experiments demonstrate that an increased\nmodel size does not consistently lead to enhanced fairness and all tested LLMs\nexhibit explicit and/or implicit gender bias, even when explicit gender\nstereotypes are absent in the inputs.", + "authors": "Xiangjue Dong, Yibo Wang, Philip S. Yu, James Caverlee", + "published": "2023-11-01", + "updated": "2023-11-01", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.00884v2", + "title": "Text classification of column headers with a controlled vocabulary: leveraging LLMs for metadata enrichment", + "abstract": "Traditional dataset retrieval systems index on metadata information rather\nthan on the data values. Thus relying primarily on manual annotations and\nhigh-quality metadata, processes known to be labour-intensive and challenging\nto automate. We propose a method to support metadata enrichment with topic\nannotations of column headers using three Large Language Models (LLMs):\nChatGPT-3.5, GoogleBard and GoogleGemini. We investigate the LLMs ability to\nclassify column headers based on domain-specific topics from a controlled\nvocabulary. We evaluate our approach by assessing the internal consistency of\nthe LLMs, the inter-machine alignment, and the human-machine agreement for the\ntopic classification task. Additionally, we investigate the impact of\ncontextual information (i.e. dataset description) on the classification\noutcomes. Our results suggest that ChatGPT and GoogleGemini outperform\nGoogleBard for internal consistency as well as LLM-human-alignment.\nInterestingly, we found that context had no impact on the LLMs performances.\nThis work proposes a novel approach that leverages LLMs for text classification\nusing a controlled topic vocabulary, which has the potential to facilitate\nautomated metadata enrichment, thereby enhancing dataset retrieval and the\nFindability, Accessibility, Interoperability and Reusability (FAIR) of research\ndata on the Web.", + "authors": "Margherita Martorana, Tobias Kuhn, Lise Stork, Jacco van Ossenbruggen", + "published": "2024-03-01", + "updated": "2024-03-05", + "primary_cat": "cs.DB", + "cats": [ + "cs.DB", + "cs.AI", + "cs.IR" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2309.03852v2", + "title": "FLM-101B: An Open LLM and How to Train It with $100K Budget", + "abstract": "Large language models (LLMs) have achieved remarkable success in NLP and\nmultimodal tasks, among others. Despite these successes, two main challenges\nremain in developing LLMs: (i) high computational cost, and (ii) fair and\nobjective evaluations. In this paper, we report a solution to significantly\nreduce LLM training cost through a growth strategy. We demonstrate that a\n101B-parameter LLM with 0.31T tokens can be trained with a budget of 100K US\ndollars. Inspired by IQ tests, we also consolidate an additional range of\nevaluations on top of existing evaluations that focus on knowledge-oriented\nabilities. These IQ evaluations include symbolic mapping, rule understanding,\npattern mining, and anti-interference. Such evaluations minimize the potential\nimpact of memorization. Experimental results show that our model, named\nFLM-101B, trained with a budget of 100K US dollars, achieves performance\ncomparable to powerful and well-known models, e.g., GPT-3 and GLM-130B,\nespecially on the additional range of IQ evaluations. The checkpoint of\nFLM-101B is released at https://huggingface.co/CofeAI/FLM-101B.", + "authors": "Xiang Li, Yiqun Yao, Xin Jiang, Xuezhi Fang, Xuying Meng, Siqi Fan, Peng Han, Jing Li, Li Du, Bowen Qin, Zheng Zhang, Aixin Sun, Yequan Wang", + "published": "2023-09-07", + "updated": "2023-09-17", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2401.15585v1", + "title": "Evaluating Gender Bias in Large Language Models via Chain-of-Thought Prompting", + "abstract": "There exist both scalable tasks, like reading comprehension and\nfact-checking, where model performance improves with model size, and unscalable\ntasks, like arithmetic reasoning and symbolic reasoning, where model\nperformance does not necessarily improve with model size. Large language models\n(LLMs) equipped with Chain-of-Thought (CoT) prompting are able to make accurate\nincremental predictions even on unscalable tasks. Unfortunately, despite their\nexceptional reasoning abilities, LLMs tend to internalize and reproduce\ndiscriminatory societal biases. Whether CoT can provide discriminatory or\negalitarian rationalizations for the implicit information in unscalable tasks\nremains an open question.\n In this study, we examine the impact of LLMs' step-by-step predictions on\ngender bias in unscalable tasks. For this purpose, we construct a benchmark for\nan unscalable task where the LLM is given a list of words comprising feminine,\nmasculine, and gendered occupational words, and is required to count the number\nof feminine and masculine words. In our CoT prompts, we require the LLM to\nexplicitly indicate whether each word in the word list is a feminine or\nmasculine before making the final predictions. With counting and handling the\nmeaning of words, this benchmark has characteristics of both arithmetic\nreasoning and symbolic reasoning. Experimental results in English show that\nwithout step-by-step prediction, most LLMs make socially biased predictions,\ndespite the task being as simple as counting words. Interestingly, CoT\nprompting reduces this unconscious social bias in LLMs and encourages fair\npredictions.", + "authors": "Masahiro Kaneko, Danushka Bollegala, Naoaki Okazaki, Timothy Baldwin", + "published": "2024-01-28", + "updated": "2024-01-28", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.09606v1", + "title": "Large Language Models and Causal Inference in Collaboration: A Comprehensive Survey", + "abstract": "Causal inference has shown potential in enhancing the predictive accuracy,\nfairness, robustness, and explainability of Natural Language Processing (NLP)\nmodels by capturing causal relationships among variables. The emergence of\ngenerative Large Language Models (LLMs) has significantly impacted various NLP\ndomains, particularly through their advanced reasoning capabilities. This\nsurvey focuses on evaluating and improving LLMs from a causal view in the\nfollowing areas: understanding and improving the LLMs' reasoning capacity,\naddressing fairness and safety issues in LLMs, complementing LLMs with\nexplanations, and handling multimodality. Meanwhile, LLMs' strong reasoning\ncapacities can in turn contribute to the field of causal inference by aiding\ncausal relationship discovery and causal effect estimations. This review\nexplores the interplay between causal inference frameworks and LLMs from both\nperspectives, emphasizing their collective potential to further the development\nof more advanced and equitable artificial intelligence systems.", + "authors": "Xiaoyu Liu, Paiheng Xu, Junda Wu, Jiaxin Yuan, Yifan Yang, Yuhang Zhou, Fuxiao Liu, Tianrui Guan, Haoliang Wang, Tong Yu, Julian McAuley, Wei Ai, Furong Huang", + "published": "2024-03-14", + "updated": "2024-03-14", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2307.11761v1", + "title": "Fairness of ChatGPT and the Role Of Explainable-Guided Prompts", + "abstract": "Our research investigates the potential of Large-scale Language Models\n(LLMs), specifically OpenAI's GPT, in credit risk assessment-a binary\nclassification task. Our findings suggest that LLMs, when directed by\njudiciously designed prompts and supplemented with domain-specific knowledge,\ncan parallel the performance of traditional Machine Learning (ML) models.\nIntriguingly, they achieve this with significantly less data-40 times less,\nutilizing merely 20 data points compared to the ML's 800. LLMs particularly\nexcel in minimizing false positives and enhancing fairness, both being vital\naspects of risk analysis. While our results did not surpass those of classical\nML models, they underscore the potential of LLMs in analogous tasks, laying a\ngroundwork for future explorations into harnessing the capabilities of LLMs in\ndiverse ML tasks.", + "authors": "Yashar Deldjoo", + "published": "2023-07-14", + "updated": "2023-07-14", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2304.03728v1", + "title": "Interpretable Unified Language Checking", + "abstract": "Despite recent concerns about undesirable behaviors generated by large\nlanguage models (LLMs), including non-factual, biased, and hateful language, we\nfind LLMs are inherent multi-task language checkers based on their latent\nrepresentations of natural and social knowledge. We present an interpretable,\nunified, language checking (UniLC) method for both human and machine-generated\nlanguage that aims to check if language input is factual and fair. While\nfairness and fact-checking tasks have been handled separately with dedicated\nmodels, we find that LLMs can achieve high performance on a combination of\nfact-checking, stereotype detection, and hate speech detection tasks with a\nsimple, few-shot, unified set of prompts. With the ``1/2-shot'' multi-task\nlanguage checking method proposed in this work, the GPT3.5-turbo model\noutperforms fully supervised baselines on several language tasks. The simple\napproach and results suggest that based on strong latent knowledge\nrepresentations, an LLM can be an adaptive and explainable tool for detecting\nmisinformation, stereotypes, and hate speech.", + "authors": "Tianhua Zhang, Hongyin Luo, Yung-Sung Chuang, Wei Fang, Luc Gaitskell, Thomas Hartvigsen, Xixin Wu, Danny Fox, Helen Meng, James Glass", + "published": "2023-04-07", + "updated": "2023-04-07", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.10199v3", + "title": "CULTURE-GEN: Revealing Global Cultural Perception in Language Models through Natural Language Prompting", + "abstract": "As the utilization of large language models (LLMs) has proliferated\nworldwide, it is crucial for them to have adequate knowledge and fair\nrepresentation for diverse global cultures. In this work, we uncover culture\nperceptions of three SOTA models on 110 countries and regions on 8\nculture-related topics through culture-conditioned generations, and extract\nsymbols from these generations that are associated to each culture by the LLM.\nWe discover that culture-conditioned generation consist of linguistic \"markers\"\nthat distinguish marginalized cultures apart from default cultures. We also\ndiscover that LLMs have an uneven degree of diversity in the culture symbols,\nand that cultures from different geographic regions have different presence in\nLLMs' culture-agnostic generation. Our findings promote further research in\nstudying the knowledge and fairness of global culture perception in LLMs. Code\nand Data can be found in: https://github.com/huihanlhh/Culture-Gen/", + "authors": "Huihan Li, Liwei Jiang, Jena D. Huang, Hyunwoo Kim, Sebastin Santy, Taylor Sorensen, Bill Yuchen Lin, Nouha Dziri, Xiang Ren, Yejin Choi", + "published": "2024-04-16", + "updated": "2024-04-26", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.08780v1", + "title": "\"Im not Racist but...\": Discovering Bias in the Internal Knowledge of Large Language Models", + "abstract": "Large language models (LLMs) have garnered significant attention for their\nremarkable performance in a continuously expanding set of natural language\nprocessing tasks. However, these models have been shown to harbor inherent\nsocietal biases, or stereotypes, which can adversely affect their performance\nin their many downstream applications. In this paper, we introduce a novel,\npurely prompt-based approach to uncover hidden stereotypes within any arbitrary\nLLM. Our approach dynamically generates a knowledge representation of internal\nstereotypes, enabling the identification of biases encoded within the LLM's\ninternal knowledge. By illuminating the biases present in LLMs and offering a\nsystematic methodology for their analysis, our work contributes to advancing\ntransparency and promoting fairness in natural language processing systems.", + "authors": "Abel Salinas, Louis Penafiel, Robert McCormack, Fred Morstatter", + "published": "2023-10-13", + "updated": "2023-10-13", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2405.02219v1", + "title": "FairEvalLLM. A Comprehensive Framework for Benchmarking Fairness in Large Language Model Recommender Systems", + "abstract": "This paper presents a framework for evaluating fairness in recommender\nsystems powered by Large Language Models (RecLLMs), addressing the need for a\nunified approach that spans various fairness dimensions including sensitivity\nto user attributes, intrinsic fairness, and discussions of fairness based on\nunderlying benefits. In addition, our framework introduces counterfactual\nevaluations and integrates diverse user group considerations to enhance the\ndiscourse on fairness evaluation for RecLLMs.\n Our key contributions include the development of a robust framework for\nfairness evaluation in LLM-based recommendations and a structured method to\ncreate \\textit{informative user profiles} from demographic data, historical\nuser preferences, and recent interactions. We argue that the latter is\nessential for enhancing personalization in such systems, especially in\ntemporal-driven scenarios. We demonstrate the utility of our framework through\npractical applications on two datasets, LastFM-1K and ML-1M. We conduct\nexperiments on a subsample of 80 users from each dataset, testing and assessing\nthe effectiveness of various prompt construction scenarios and in-context\nlearning, comprising more than 50 scenarios. This results in more than 4000\nrecommendations (80 * 50 = 4000). Our study reveals that while there are no\nsignificant unfairness issues in scenarios involving sensitive attributes, some\nconcerns remain. However, in terms of intrinsic fairness, which does not\ninvolve direct sensitivity, unfairness across demographic groups remains\nsignificant. The code and data used for this paper are available at:\n\\url{https://shorturl.at/awBFM}.", + "authors": "Yashar Deldjoo", + "published": "2024-05-03", + "updated": "2024-05-03", + "primary_cat": "cs.IR", + "cats": [ + "cs.IR" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2312.07420v1", + "title": "FairSISA: Ensemble Post-Processing to Improve Fairness of Unlearning in LLMs", + "abstract": "Training large language models (LLMs) is a costly endeavour in terms of time\nand computational resources. The large amount of training data used during the\nunsupervised pre-training phase makes it difficult to verify all data and,\nunfortunately, undesirable data may be ingested during training. Re-training\nfrom scratch is impractical and has led to the creation of the 'unlearning'\ndiscipline where models are modified to \"unlearn\" undesirable information\nwithout retraining. However, any modification can alter the behaviour of LLMs,\nespecially on key dimensions such as fairness. This is the first work that\nexamines this interplay between unlearning and fairness for LLMs. In\nparticular, we focus on a popular unlearning framework known as SISA [Bourtoule\net al., 2021], which creates an ensemble of models trained on disjoint shards.\nWe evaluate the performance-fairness trade-off for SISA, and empirically\ndemsontrate that SISA can indeed reduce fairness in LLMs. To remedy this, we\npropose post-processing bias mitigation techniques for ensemble models produced\nby SISA. We adapt the post-processing fairness improvement technique from\n[Hardt et al., 2016] to design three methods that can handle model ensembles,\nand prove that one of the methods is an optimal fair predictor for ensemble of\nmodels. Through experimental results, we demonstrate the efficacy of our\npost-processing framework called 'FairSISA'.", + "authors": "Swanand Ravindra Kadhe, Anisa Halimi, Ambrish Rawat, Nathalie Baracaldo", + "published": "2023-12-12", + "updated": "2023-12-12", + "primary_cat": "cs.LG", + "cats": [ + "cs.LG", + "cs.CY" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.06852v2", + "title": "ChemLLM: A Chemical Large Language Model", + "abstract": "Large language models (LLMs) have made impressive progress in chemistry\napplications. However, the community lacks an LLM specifically designed for\nchemistry. The main challenges are two-fold: firstly, most chemical data and\nscientific knowledge are stored in structured databases, which limits the\nmodel's ability to sustain coherent dialogue when used directly. Secondly,\nthere is an absence of objective and fair benchmark that encompass most\nchemistry tasks. Here, we introduce ChemLLM, a comprehensive framework that\nfeatures the first LLM dedicated to chemistry. It also includes ChemData, a\ndataset specifically designed for instruction tuning, and ChemBench, a robust\nbenchmark covering nine essential chemistry tasks. ChemLLM is adept at\nperforming various tasks across chemical disciplines with fluid dialogue\ninteraction. Notably, ChemLLM achieves results comparable to GPT-4 on the core\nchemical tasks and demonstrates competitive performance with LLMs of similar\nsize in general scenarios. ChemLLM paves a new path for exploration in chemical\nstudies, and our method of incorporating structured chemical knowledge into\ndialogue systems sets a new standard for developing LLMs in various scientific\nfields. Codes, Datasets, and Model weights are publicly accessible at\nhttps://hf.co/AI4Chem", + "authors": "Di Zhang, Wei Liu, Qian Tan, Jingdan Chen, Hang Yan, Yuliang Yan, Jiatong Li, Weiran Huang, Xiangyu Yue, Wanli Ouyang, Dongzhan Zhou, Shufei Zhang, Mao Su, Han-Sen Zhong, Yuqiang Li", + "published": "2024-02-10", + "updated": "2024-04-25", + "primary_cat": "cs.AI", + "cats": [ + "cs.AI", + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2307.15997v1", + "title": "RoCar: A Relationship Network-based Evaluation Method to Large Language Models", + "abstract": "Large language models (LLMs) have received increasing attention. However, due\nto the complexity of its capabilities, how to rationally evaluate the\ncapabilities of LLMs is still a task to be solved. We propose the RoCar method,\nwhich utilizes the defined basic schemas to randomly construct a task graph and\ngenerates natural language evaluation tasks based on the task graph to evaluate\nthe reasoning and memory abilities of LLMs respectively. Due to the very large\nrandomness of the task construction process, it is possible to ensure that none\nof the LLMs to be tested has directly learned the evaluation tasks,\nguaranteeing the fairness of the evaluation method.", + "authors": "Ming Wang, Wenfang Wu, Chongyun Gao, Daling Wang, Shi Feng, Yifei Zhang", + "published": "2023-07-29", + "updated": "2023-07-29", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2308.05374v2", + "title": "Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignment", + "abstract": "Ensuring alignment, which refers to making models behave in accordance with\nhuman intentions [1,2], has become a critical task before deploying large\nlanguage models (LLMs) in real-world applications. For instance, OpenAI devoted\nsix months to iteratively aligning GPT-4 before its release [3]. However, a\nmajor challenge faced by practitioners is the lack of clear guidance on\nevaluating whether LLM outputs align with social norms, values, and\nregulations. This obstacle hinders systematic iteration and deployment of LLMs.\nTo address this issue, this paper presents a comprehensive survey of key\ndimensions that are crucial to consider when assessing LLM trustworthiness. The\nsurvey covers seven major categories of LLM trustworthiness: reliability,\nsafety, fairness, resistance to misuse, explainability and reasoning, adherence\nto social norms, and robustness. Each major category is further divided into\nseveral sub-categories, resulting in a total of 29 sub-categories.\nAdditionally, a subset of 8 sub-categories is selected for further\ninvestigation, where corresponding measurement studies are designed and\nconducted on several widely-used LLMs. The measurement results indicate that,\nin general, more aligned models tend to perform better in terms of overall\ntrustworthiness. However, the effectiveness of alignment varies across the\ndifferent trustworthiness categories considered. This highlights the importance\nof conducting more fine-grained analyses, testing, and making continuous\nimprovements on LLM alignment. By shedding light on these key dimensions of LLM\ntrustworthiness, this paper aims to provide valuable insights and guidance to\npractitioners in the field. Understanding and addressing these concerns will be\ncrucial in achieving reliable and ethically sound deployment of LLMs in various\napplications.", + "authors": "Yang Liu, Yuanshun Yao, Jean-Francois Ton, Xiaoying Zhang, Ruocheng Guo, Hao Cheng, Yegor Klochkov, Muhammad Faaiz Taufiq, Hang Li", + "published": "2023-08-10", + "updated": "2024-03-21", + "primary_cat": "cs.AI", + "cats": [ + "cs.AI", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.14208v2", + "title": "Content Conditional Debiasing for Fair Text Embedding", + "abstract": "Mitigating biases in machine learning models has gained increasing attention\nin Natural Language Processing (NLP). Yet, only a few studies focus on fair\ntext embeddings, which are crucial yet challenging for real-world applications.\nIn this paper, we propose a novel method for learning fair text embeddings. We\nachieve fairness while maintaining utility trade-off by ensuring conditional\nindependence between sensitive attributes and text embeddings conditioned on\nthe content. Specifically, we enforce that embeddings of texts with different\nsensitive attributes but identical content maintain the same distance toward\nthe embedding of their corresponding neutral text. Furthermore, we address the\nissue of lacking proper training data by using Large Language Models (LLMs) to\naugment texts into different sensitive groups. Our extensive evaluations\ndemonstrate that our approach effectively improves fairness while preserving\nthe utility of embeddings, representing a pioneering effort in achieving\nconditional independence for fair text embeddings.", + "authors": "Wenlong Deng, Blair Chen, Xiaoxiao Li, Christos Thrampoulidis", + "published": "2024-02-22", + "updated": "2024-02-23", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.CY", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.12736v1", + "title": "Large Language Model Supply Chain: A Research Agenda", + "abstract": "The rapid advancements in pre-trained Large Language Models (LLMs) and Large\nMultimodal Models (LMMs) have ushered in a new era of intelligent applications,\ntransforming fields ranging from natural language processing to content\ngeneration. The LLM supply chain represents a crucial aspect of the\ncontemporary artificial intelligence landscape. It encompasses the entire\nlifecycle of pre-trained models, from its initial development and training to\nits final deployment and application in various domains. This paper presents a\ncomprehensive overview of the LLM supply chain, highlighting its three core\nelements: 1) the model infrastructure, encompassing datasets and toolchain for\ntraining, optimization, and deployment; 2) the model lifecycle, covering\ntraining, testing, releasing, and ongoing maintenance; and 3) the downstream\napplication ecosystem, enabling the integration of pre-trained models into a\nwide range of intelligent applications. However, this rapidly evolving field\nfaces numerous challenges across these key components, including data privacy\nand security, model interpretability and fairness, infrastructure scalability,\nand regulatory compliance. Addressing these challenges is essential for\nharnessing the full potential of LLMs and ensuring their ethical and\nresponsible use. This paper provides a future research agenda for the LLM\nsupply chain, aiming at driving the continued advancement and responsible\ndeployment of these transformative LLMs.", + "authors": "Shenao Wang, Yanjie Zhao, Xinyi Hou, Haoyu Wang", + "published": "2024-04-19", + "updated": "2024-04-19", + "primary_cat": "cs.SE", + "cats": [ + "cs.SE" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.16343v2", + "title": "Evaluating, Understanding, and Improving Constrained Text Generation for Large Language Models", + "abstract": "Advancements in natural language generation (NLG) and large language models\n(LLMs) have led to proficient text generation in various tasks. However,\nintegrating intricate constraints into neural text generation, due to LLMs'\nopacity, remains challenging. This study investigates constrained text\ngeneration for LLMs, where predefined constraints are applied during LLM's\ngeneration process. Our research mainly focuses on mainstream open-source LLMs,\ncategorizing constraints into lexical, structural, and relation-based types. We\nalso present various benchmarks to facilitate fair evaluation. The study\naddresses some key research questions, including evaluating, understanding and\nimproving constrained text generation for LLMs. Results illuminate LLMs'\ncapacity and deficiency to incorporate constraints and provide insights for\nfuture developments in constrained text generation. Codes and datasets will be\nreleased upon acceptance.", + "authors": "Xiang Chen, Xiaojun Wan", + "published": "2023-10-25", + "updated": "2024-03-21", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2312.15478v1", + "title": "A Group Fairness Lens for Large Language Models", + "abstract": "The rapid advancement of large language models has revolutionized various\napplications but also raised crucial concerns about their potential to\nperpetuate biases and unfairness when deployed in social media contexts.\nEvaluating LLMs' potential biases and fairness has become crucial, as existing\nmethods rely on limited prompts focusing on just a few groups, lacking a\ncomprehensive categorical perspective. In this paper, we propose evaluating LLM\nbiases from a group fairness lens using a novel hierarchical schema\ncharacterizing diverse social groups. Specifically, we construct a dataset,\nGFair, encapsulating target-attribute combinations across multiple dimensions.\nIn addition, we introduce statement organization, a new open-ended text\ngeneration task, to uncover complex biases in LLMs. Extensive evaluations of\npopular LLMs reveal inherent safety concerns. To mitigate the biases of LLM\nfrom a group fairness perspective, we pioneer a novel chain-of-thought method\nGF-Think to mitigate biases of LLMs from a group fairness perspective.\nExperimental results demonstrate its efficacy in mitigating bias in LLMs to\nachieve fairness.", + "authors": "Guanqun Bi, Lei Shen, Yuqiang Xie, Yanan Cao, Tiangang Zhu, Xiaodong He", + "published": "2023-12-24", + "updated": "2023-12-24", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.13925v1", + "title": "MARIO Eval: Evaluate Your Math LLM with your Math LLM--A mathematical dataset evaluation toolkit", + "abstract": "Large language models (LLMs) have been explored in a variety of reasoning\ntasks including solving of mathematical problems. Each math dataset typically\nincludes its own specially designed evaluation script, which, while suitable\nfor its intended use, lacks generalizability across different datasets.\nConsequently, updates and adaptations to these evaluation tools tend to occur\nwithout being systematically reported, leading to inconsistencies and obstacles\nto fair comparison across studies. To bridge this gap, we introduce a\ncomprehensive mathematical evaluation toolkit that not only utilizes a python\ncomputer algebra system (CAS) for its numerical accuracy, but also integrates\nan optional LLM, known for its considerable natural language processing\ncapabilities. To validate the effectiveness of our toolkit, we manually\nannotated two distinct datasets. Our experiments demonstrate that the toolkit\nyields more robust evaluation results compared to prior works, even without an\nLLM. Furthermore, when an LLM is incorporated, there is a notable enhancement.\nThe code for our method will be made available at\n\\url{https://github.com/MARIO-Math-Reasoning/math_evaluation}.", + "authors": "Boning Zhang, Chengxi Li, Kai Fan", + "published": "2024-04-22", + "updated": "2024-04-22", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.06003v1", + "title": "FreeEval: A Modular Framework for Trustworthy and Efficient Evaluation of Large Language Models", + "abstract": "The rapid development of large language model (LLM) evaluation methodologies\nand datasets has led to a profound challenge: integrating state-of-the-art\nevaluation techniques cost-effectively while ensuring reliability,\nreproducibility, and efficiency. Currently, there is a notable absence of a\nunified and adaptable framework that seamlessly integrates various evaluation\napproaches. Moreover, the reliability of evaluation findings is often\nquestionable due to potential data contamination, with the evaluation\nefficiency commonly overlooked when facing the substantial costs associated\nwith LLM inference. In response to these challenges, we introduce FreeEval, a\nmodular and scalable framework crafted to enable trustworthy and efficient\nautomatic evaluations of LLMs. Firstly, FreeEval's unified abstractions\nsimplify the integration and improve the transparency of diverse evaluation\nmethodologies, encompassing dynamic evaluation that demand sophisticated LLM\ninteractions. Secondly, the framework integrates meta-evaluation techniques\nlike human evaluation and data contamination detection, which, along with\ndynamic evaluation modules in the platform, enhance the fairness of the\nevaluation outcomes. Lastly, FreeEval is designed with a high-performance\ninfrastructure, including distributed computation and caching strategies,\nenabling extensive evaluations across multi-node, multi-GPU clusters for\nopen-source and proprietary LLMs.", + "authors": "Zhuohao Yu, Chang Gao, Wenjin Yao, Yidong Wang, Zhengran Zeng, Wei Ye, Jindong Wang, Yue Zhang, Shikun Zhang", + "published": "2024-04-09", + "updated": "2024-04-09", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2305.11595v3", + "title": "Examining Inter-Consistency of Large Language Models Collaboration: An In-depth Analysis via Debate", + "abstract": "Large Language Models (LLMs) have shown impressive capabilities in various\napplications, but they still face various inconsistency issues. Existing works\nprimarily focus on the inconsistency issues within a single LLM, while we\ncomplementarily explore the inter-consistency among multiple LLMs for\ncollaboration. To examine whether LLMs can collaborate effectively to achieve a\nconsensus for a shared goal, we focus on commonsense reasoning, and introduce a\nformal debate framework (FORD) to conduct a three-stage debate among LLMs with\nreal-world scenarios alignment: fair debate, mismatched debate, and roundtable\ndebate. Through extensive experiments on various datasets, LLMs can effectively\ncollaborate to reach a consensus despite noticeable inter-inconsistencies, but\nimbalances in their abilities can lead to domination by superior LLMs.\nLeveraging a more advanced LLM like GPT-4 as an authoritative judge can boost\ncollaboration performance. Our work contributes to understanding the\ninter-consistency among LLMs and lays the foundation for developing future\ncollaboration methods. Codes and data are available at\nhttps://github.com/Waste-Wood/FORD", + "authors": "Kai Xiong, Xiao Ding, Yixin Cao, Ting Liu, Bing Qin", + "published": "2023-05-19", + "updated": "2023-10-18", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.07981v1", + "title": "Manipulating Large Language Models to Increase Product Visibility", + "abstract": "Large language models (LLMs) are increasingly being integrated into search\nengines to provide natural language responses tailored to user queries.\nCustomers and end-users are also becoming more dependent on these models for\nquick and easy purchase decisions. In this work, we investigate whether\nrecommendations from LLMs can be manipulated to enhance a product's visibility.\nWe demonstrate that adding a strategic text sequence (STS) -- a carefully\ncrafted message -- to a product's information page can significantly increase\nits likelihood of being listed as the LLM's top recommendation. To understand\nthe impact of STS, we use a catalog of fictitious coffee machines and analyze\nits effect on two target products: one that seldom appears in the LLM's\nrecommendations and another that usually ranks second. We observe that the\nstrategic text sequence significantly enhances the visibility of both products\nby increasing their chances of appearing as the top recommendation. This\nability to manipulate LLM-generated search responses provides vendors with a\nconsiderable competitive advantage and has the potential to disrupt fair market\ncompetition. Just as search engine optimization (SEO) revolutionized how\nwebpages are customized to rank higher in search engine results, influencing\nLLM recommendations could profoundly impact content optimization for AI-driven\nsearch services. Code for our experiments is available at\nhttps://github.com/aounon/llm-rank-optimizer.", + "authors": "Aounon Kumar, Himabindu Lakkaraju", + "published": "2024-04-11", + "updated": "2024-04-11", + "primary_cat": "cs.IR", + "cats": [ + "cs.IR", + "cs.AI", + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.17553v1", + "title": "RuBia: A Russian Language Bias Detection Dataset", + "abstract": "Warning: this work contains upsetting or disturbing content.\n Large language models (LLMs) tend to learn the social and cultural biases\npresent in the raw pre-training data. To test if an LLM's behavior is fair,\nfunctional datasets are employed, and due to their purpose, these datasets are\nhighly language and culture-specific. In this paper, we address a gap in the\nscope of multilingual bias evaluation by presenting a bias detection dataset\nspecifically designed for the Russian language, dubbed as RuBia. The RuBia\ndataset is divided into 4 domains: gender, nationality, socio-economic status,\nand diverse, each of the domains is further divided into multiple fine-grained\nsubdomains. Every example in the dataset consists of two sentences with the\nfirst reinforcing a potentially harmful stereotype or trope and the second\ncontradicting it. These sentence pairs were first written by volunteers and\nthen validated by native-speaking crowdsourcing workers. Overall, there are\nnearly 2,000 unique sentence pairs spread over 19 subdomains in RuBia. To\nillustrate the dataset's purpose, we conduct a diagnostic evaluation of\nstate-of-the-art or near-state-of-the-art LLMs and discuss the LLMs'\npredisposition to social biases.", + "authors": "Veronika Grigoreva, Anastasiia Ivanova, Ilseyar Alimova, Ekaterina Artemova", + "published": "2024-03-26", + "updated": "2024-03-26", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.00811v1", + "title": "Cognitive Bias in High-Stakes Decision-Making with LLMs", + "abstract": "Large language models (LLMs) offer significant potential as tools to support\nan expanding range of decision-making tasks. However, given their training on\nhuman (created) data, LLMs can inherit both societal biases against protected\ngroups, as well as be subject to cognitive bias. Such human-like bias can\nimpede fair and explainable decisions made with LLM assistance. Our work\nintroduces BiasBuster, a framework designed to uncover, evaluate, and mitigate\ncognitive bias in LLMs, particularly in high-stakes decision-making tasks.\nInspired by prior research in psychology and cognitive sciences, we develop a\ndataset containing 16,800 prompts to evaluate different cognitive biases (e.g.,\nprompt-induced, sequential, inherent). We test various bias mitigation\nstrategies, amidst proposing a novel method using LLMs to debias their own\nprompts. Our analysis provides a comprehensive picture on the presence and\neffects of cognitive bias across different commercial and open-source models.\nWe demonstrate that our self-help debiasing effectively mitigate cognitive bias\nwithout having to manually craft examples for each bias type.", + "authors": "Jessica Echterhoff, Yao Liu, Abeer Alessa, Julian McAuley, Zexue He", + "published": "2024-02-25", + "updated": "2024-02-25", + "primary_cat": "cs.AI", + "cats": [ + "cs.AI", + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.09219v5", + "title": "\"Kelly is a Warm Person, Joseph is a Role Model\": Gender Biases in LLM-Generated Reference Letters", + "abstract": "Large Language Models (LLMs) have recently emerged as an effective tool to\nassist individuals in writing various types of content, including professional\ndocuments such as recommendation letters. Though bringing convenience, this\napplication also introduces unprecedented fairness concerns. Model-generated\nreference letters might be directly used by users in professional scenarios. If\nunderlying biases exist in these model-constructed letters, using them without\nscrutinization could lead to direct societal harms, such as sabotaging\napplication success rates for female applicants. In light of this pressing\nissue, it is imminent and necessary to comprehensively study fairness issues\nand associated harms in this real-world use case. In this paper, we critically\nexamine gender biases in LLM-generated reference letters. Drawing inspiration\nfrom social science findings, we design evaluation methods to manifest biases\nthrough 2 dimensions: (1) biases in language style and (2) biases in lexical\ncontent. We further investigate the extent of bias propagation by analyzing the\nhallucination bias of models, a term that we define to be bias exacerbation in\nmodel-hallucinated contents. Through benchmarking evaluation on 2 popular LLMs-\nChatGPT and Alpaca, we reveal significant gender biases in LLM-generated\nrecommendation letters. Our findings not only warn against using LLMs for this\napplication without scrutinization, but also illuminate the importance of\nthoroughly studying hidden biases and harms in LLM-generated professional\ndocuments.", + "authors": "Yixin Wan, George Pu, Jiao Sun, Aparna Garimella, Kai-Wei Chang, Nanyun Peng", + "published": "2023-10-13", + "updated": "2023-12-01", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.01964v1", + "title": "Don't Make Your LLM an Evaluation Benchmark Cheater", + "abstract": "Large language models~(LLMs) have greatly advanced the frontiers of\nartificial intelligence, attaining remarkable improvement in model capacity. To\nassess the model performance, a typical approach is to construct evaluation\nbenchmarks for measuring the ability level of LLMs in different aspects.\nDespite that a number of high-quality benchmarks have been released, the\nconcerns about the appropriate use of these benchmarks and the fair comparison\nof different models are increasingly growing. Considering these concerns, in\nthis paper, we discuss the potential risk and impact of inappropriately using\nevaluation benchmarks and misleadingly interpreting the evaluation results.\nSpecially, we focus on a special issue that would lead to inappropriate\nevaluation, \\ie \\emph{benchmark leakage}, referring that the data related to\nevaluation sets is occasionally used for model training. This phenomenon now\nbecomes more common since pre-training data is often prepared ahead of model\ntest. We conduct extensive experiments to study the effect of benchmark\nleverage, and find that it can dramatically boost the evaluation results, which\nwould finally lead to an unreliable assessment of model performance. To improve\nthe use of existing evaluation benchmarks, we finally present several\nguidelines for both LLM developers and benchmark maintainers. We hope this work\ncan draw attention to appropriate training and evaluation of LLMs.", + "authors": "Kun Zhou, Yutao Zhu, Zhipeng Chen, Wentong Chen, Wayne Xin Zhao, Xu Chen, Yankai Lin, Ji-Rong Wen, Jiawei Han", + "published": "2023-11-03", + "updated": "2023-11-03", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.18580v1", + "title": "FFT: Towards Harmlessness Evaluation and Analysis for LLMs with Factuality, Fairness, Toxicity", + "abstract": "The widespread of generative artificial intelligence has heightened concerns\nabout the potential harms posed by AI-generated texts, primarily stemming from\nfactoid, unfair, and toxic content. Previous researchers have invested much\neffort in assessing the harmlessness of generative language models. However,\nexisting benchmarks are struggling in the era of large language models (LLMs),\ndue to the stronger language generation and instruction following capabilities,\nas well as wider applications. In this paper, we propose FFT, a new benchmark\nwith 2116 elaborated-designed instances, for LLM harmlessness evaluation with\nfactuality, fairness, and toxicity. To investigate the potential harms of LLMs,\nwe evaluate 9 representative LLMs covering various parameter scales, training\nstages, and creators. Experiments show that the harmlessness of LLMs is still\nunder-satisfactory, and extensive analysis derives some insightful findings\nthat could inspire future research for harmless LLM research.", + "authors": "Shiyao Cui, Zhenyu Zhang, Yilong Chen, Wenyuan Zhang, Tianyun Liu, Siqi Wang, Tingwen Liu", + "published": "2023-11-30", + "updated": "2023-11-30", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.CR" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.02049v1", + "title": "Post Turing: Mapping the landscape of LLM Evaluation", + "abstract": "In the rapidly evolving landscape of Large Language Models (LLMs),\nintroduction of well-defined and standardized evaluation methodologies remains\na crucial challenge. This paper traces the historical trajectory of LLM\nevaluations, from the foundational questions posed by Alan Turing to the modern\nera of AI research. We categorize the evolution of LLMs into distinct periods,\neach characterized by its unique benchmarks and evaluation criteria. As LLMs\nincreasingly mimic human-like behaviors, traditional evaluation proxies, such\nas the Turing test, have become less reliable. We emphasize the pressing need\nfor a unified evaluation system, given the broader societal implications of\nthese models. Through an analysis of common evaluation methodologies, we\nadvocate for a qualitative shift in assessment approaches, underscoring the\nimportance of standardization and objective criteria. This work serves as a\ncall for the AI community to collaboratively address the challenges of LLM\nevaluation, ensuring their reliability, fairness, and societal benefit.", + "authors": "Alexey Tikhonov, Ivan P. Yamshchikov", + "published": "2023-11-03", + "updated": "2023-11-03", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "68T50", + "I.2.7" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2309.11653v2", + "title": "\"It's a Fair Game\", or Is It? Examining How Users Navigate Disclosure Risks and Benefits When Using LLM-Based Conversational Agents", + "abstract": "The widespread use of Large Language Model (LLM)-based conversational agents\n(CAs), especially in high-stakes domains, raises many privacy concerns.\nBuilding ethical LLM-based CAs that respect user privacy requires an in-depth\nunderstanding of the privacy risks that concern users the most. However,\nexisting research, primarily model-centered, does not provide insight into\nusers' perspectives. To bridge this gap, we analyzed sensitive disclosures in\nreal-world ChatGPT conversations and conducted semi-structured interviews with\n19 LLM-based CA users. We found that users are constantly faced with trade-offs\nbetween privacy, utility, and convenience when using LLM-based CAs. However,\nusers' erroneous mental models and the dark patterns in system design limited\ntheir awareness and comprehension of the privacy risks. Additionally, the\nhuman-like interactions encouraged more sensitive disclosures, which\ncomplicated users' ability to navigate the trade-offs. We discuss practical\ndesign guidelines and the needs for paradigm shifts to protect the privacy of\nLLM-based CA users.", + "authors": "Zhiping Zhang, Michelle Jia, Hao-Ping Lee, Bingsheng Yao, Sauvik Das, Ada Lerner, Dakuo Wang, Tianshi Li", + "published": "2023-09-20", + "updated": "2024-04-02", + "primary_cat": "cs.HC", + "cats": [ + "cs.HC", + "cs.AI", + "cs.CR" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2303.01248v3", + "title": "Can ChatGPT Assess Human Personalities? A General Evaluation Framework", + "abstract": "Large Language Models (LLMs) especially ChatGPT have produced impressive\nresults in various areas, but their potential human-like psychology is still\nlargely unexplored. Existing works study the virtual personalities of LLMs but\nrarely explore the possibility of analyzing human personalities via LLMs. This\npaper presents a generic evaluation framework for LLMs to assess human\npersonalities based on Myers Briggs Type Indicator (MBTI) tests. Specifically,\nwe first devise unbiased prompts by randomly permuting options in MBTI\nquestions and adopt the average testing result to encourage more impartial\nanswer generation. Then, we propose to replace the subject in question\nstatements to enable flexible queries and assessments on different subjects\nfrom LLMs. Finally, we re-formulate the question instructions in a manner of\ncorrectness evaluation to facilitate LLMs to generate clearer responses. The\nproposed framework enables LLMs to flexibly assess personalities of different\ngroups of people. We further propose three evaluation metrics to measure the\nconsistency, robustness, and fairness of assessment results from\nstate-of-the-art LLMs including ChatGPT and GPT-4. Our experiments reveal\nChatGPT's ability to assess human personalities, and the average results\ndemonstrate that it can achieve more consistent and fairer assessments in spite\nof lower robustness against prompt biases compared with InstructGPT.", + "authors": "Haocong Rao, Cyril Leung, Chunyan Miao", + "published": "2023-03-01", + "updated": "2023-10-13", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.02680v1", + "title": "Large Language Models are Geographically Biased", + "abstract": "Large Language Models (LLMs) inherently carry the biases contained in their\ntraining corpora, which can lead to the perpetuation of societal harm. As the\nimpact of these foundation models grows, understanding and evaluating their\nbiases becomes crucial to achieving fairness and accuracy. We propose to study\nwhat LLMs know about the world we live in through the lens of geography. This\napproach is particularly powerful as there is ground truth for the numerous\naspects of human life that are meaningfully projected onto geographic space\nsuch as culture, race, language, politics, and religion. We show various\nproblematic geographic biases, which we define as systemic errors in geospatial\npredictions. Initially, we demonstrate that LLMs are capable of making accurate\nzero-shot geospatial predictions in the form of ratings that show strong\nmonotonic correlation with ground truth (Spearman's $\\rho$ of up to 0.89). We\nthen show that LLMs exhibit common biases across a range of objective and\nsubjective topics. In particular, LLMs are clearly biased against locations\nwith lower socioeconomic conditions (e.g. most of Africa) on a variety of\nsensitive subjective topics such as attractiveness, morality, and intelligence\n(Spearman's $\\rho$ of up to 0.70). Finally, we introduce a bias score to\nquantify this and find that there is significant variation in the magnitude of\nbias across existing LLMs.", + "authors": "Rohin Manvi, Samar Khanna, Marshall Burke, David Lobell, Stefano Ermon", + "published": "2024-02-05", + "updated": "2024-02-05", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.CY", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.18276v1", + "title": "Bias Neutralization Framework: Measuring Fairness in Large Language Models with Bias Intelligence Quotient (BiQ)", + "abstract": "The burgeoning influence of Large Language Models (LLMs) in shaping public\ndiscourse and decision-making underscores the imperative to address inherent\nbiases within these AI systems. In the wake of AI's expansive integration\nacross sectors, addressing racial bias in LLMs has never been more critical.\nThis paper introduces a novel framework called Comprehensive Bias\nNeutralization Framework (CBNF) which embodies an innovative approach to\nquantifying and mitigating biases within LLMs. Our framework combines the Large\nLanguage Model Bias Index (LLMBI) [Oketunji, A., Anas, M., Saina, D., (2023)]\nand Bias removaL with No Demographics (BLIND) [Orgad, H., Belinkov, Y. (2023)]\nmethodologies to create a new metric called Bias Intelligence Quotient\n(BiQ)which detects, measures, and mitigates racial bias in LLMs without\nreliance on demographic annotations.\n By introducing a new metric called BiQ that enhances LLMBI with additional\nfairness metrics, CBNF offers a multi-dimensional metric for bias assessment,\nunderscoring the necessity of a nuanced approach to fairness in AI [Mehrabi et\nal., 2021]. This paper presents a detailed analysis of Latimer AI (a language\nmodel incrementally trained on black history and culture) in comparison to\nChatGPT 3.5, illustrating Latimer AI's efficacy in detecting racial, cultural,\nand gender biases through targeted training and refined bias mitigation\nstrategies [Latimer & Bender, 2023].", + "authors": "Malur Narayan, John Pasmore, Elton Sampaio, Vijay Raghavan, Gabriella Waters", + "published": "2024-04-28", + "updated": "2024-04-28", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "D.1; I.2" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2309.14345v2", + "title": "Bias Testing and Mitigation in LLM-based Code Generation", + "abstract": "Utilizing state-of-the-art Large Language Models (LLMs), automatic code\ngeneration models play a pivotal role in enhancing the productivity of software\ndevelopment procedures. As the adoption of LLMs becomes more widespread in\nsoftware coding ecosystems, a pressing issue has emerged: does the generated\ncode contain social bias and unfairness, such as those related to age, gender,\nand race? This issue concerns the integrity, fairness, and ethical foundation\nof software applications that depend on the code generated by these models, yet\nis under-explored in the literature. This paper presents a novel bias testing\nframework that is specifically designed for code generation tasks. Based on\nthis framework, we conduct an extensive evaluation of the bias in code\ngenerated by five state-of-the-art LLMs. Our findings reveal that 20.29% to\n44.93% code functions generated by the models under study are biased when\nhandling bias sensitive tasks (i.e., tasks that involve sensitive attributes\nsuch as age and gender). This indicates that the existing LLMs can be unfair in\ncode generation, posing risks of unintended and harmful software behaviors. To\nmitigate bias for code generation models, we evaluate five bias mitigation\nprompt strategies, i.e., utilizing bias testing results to refine the code\n(zero-shot), one-, few-shot, and two Chain-of-Thought (CoT) prompts. Our\nevaluation results illustrate that these strategies are all effective in\nmitigating bias. Overall, one-shot and few-shot learning are the two most\neffective. For GPT-4, 80% to 90% code bias can be removed with one-shot\nlearning.", + "authors": "Dong Huang, Qingwen Bu, Jie Zhang, Xiaofei Xie, Junjie Chen, Heming Cui", + "published": "2023-09-03", + "updated": "2024-01-09", + "primary_cat": "cs.SE", + "cats": [ + "cs.SE", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.02294v1", + "title": "LLMs grasp morality in concept", + "abstract": "Work in AI ethics and fairness has made much progress in regulating LLMs to\nreflect certain values, such as fairness, truth, and diversity. However, it has\ntaken the problem of how LLMs might 'mean' anything at all for granted. Without\naddressing this, it is not clear what imbuing LLMs with such values even means.\nIn response, we provide a general theory of meaning that extends beyond humans.\nWe use this theory to explicate the precise nature of LLMs as meaning-agents.\nWe suggest that the LLM, by virtue of its position as a meaning-agent, already\ngrasps the constructions of human society (e.g. morality, gender, and race) in\nconcept. Consequently, under certain ethical frameworks, currently popular\nmethods for model alignment are limited at best and counterproductive at worst.\nMoreover, unaligned models may help us better develop our moral and social\nphilosophy.", + "authors": "Mark Pock, Andre Ye, Jared Moore", + "published": "2023-11-04", + "updated": "2023-11-04", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.CY" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.06500v1", + "title": "MetaAgents: Simulating Interactions of Human Behaviors for LLM-based Task-oriented Coordination via Collaborative Generative Agents", + "abstract": "Significant advancements have occurred in the application of Large Language\nModels (LLMs) for various tasks and social simulations. Despite this, their\ncapacities to coordinate within task-oriented social contexts are\nunder-explored. Such capabilities are crucial if LLMs are to effectively mimic\nhuman-like social behavior and produce meaningful results. To bridge this gap,\nwe introduce collaborative generative agents, endowing LLM-based Agents with\nconsistent behavior patterns and task-solving abilities. We situate these\nagents in a simulated job fair environment as a case study to scrutinize their\ncoordination skills. We propose a novel framework that equips collaborative\ngenerative agents with human-like reasoning abilities and specialized skills.\nOur evaluation demonstrates that these agents show promising performance.\nHowever, we also uncover limitations that hinder their effectiveness in more\ncomplex coordination tasks. Our work provides valuable insights into the role\nand evolution of LLMs in task-oriented social simulations.", + "authors": "Yuan Li, Yixuan Zhang, Lichao Sun", + "published": "2023-10-10", + "updated": "2023-10-10", + "primary_cat": "cs.AI", + "cats": [ + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.02650v1", + "title": "Towards detecting unanticipated bias in Large Language Models", + "abstract": "Over the last year, Large Language Models (LLMs) like ChatGPT have become\nwidely available and have exhibited fairness issues similar to those in\nprevious machine learning systems. Current research is primarily focused on\nanalyzing and quantifying these biases in training data and their impact on the\ndecisions of these models, alongside developing mitigation strategies. This\nresearch largely targets well-known biases related to gender, race, ethnicity,\nand language. However, it is clear that LLMs are also affected by other, less\nobvious implicit biases. The complex and often opaque nature of these models\nmakes detecting such biases challenging, yet this is crucial due to their\npotential negative impact in various applications. In this paper, we explore\nnew avenues for detecting these unanticipated biases in LLMs, focusing\nspecifically on Uncertainty Quantification and Explainable AI methods. These\napproaches aim to assess the certainty of model decisions and to make the\ninternal decision-making processes of LLMs more transparent, thereby\nidentifying and understanding biases that are not immediately apparent. Through\nthis research, we aim to contribute to the development of fairer and more\ntransparent AI systems.", + "authors": "Anna Kruspe", + "published": "2024-04-03", + "updated": "2024-04-03", + "primary_cat": "cs.LG", + "cats": [ + "cs.LG", + "cs.AI", + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.18140v1", + "title": "ROBBIE: Robust Bias Evaluation of Large Generative Language Models", + "abstract": "As generative large language models (LLMs) grow more performant and\nprevalent, we must develop comprehensive enough tools to measure and improve\ntheir fairness. Different prompt-based datasets can be used to measure social\nbias across multiple text domains and demographic axes, meaning that testing\nLLMs on more datasets can potentially help us characterize their biases more\nfully, and better ensure equal and equitable treatment of marginalized\ndemographic groups. In this work, our focus is two-fold:\n (1) Benchmarking: a comparison of 6 different prompt-based bias and toxicity\nmetrics across 12 demographic axes and 5 families of generative LLMs. Out of\nthose 6 metrics, AdvPromptSet and HolisticBiasR are novel datasets proposed in\nthe paper. The comparison of those benchmarks gives us insights about the bias\nand toxicity of the compared models. Therefore, we explore the frequency of\ndemographic terms in common LLM pre-training corpora and how this may relate to\nmodel biases.\n (2) Mitigation: we conduct a comprehensive study of how well 3 bias/toxicity\nmitigation techniques perform across our suite of measurements. ROBBIE aims to\nprovide insights for practitioners while deploying a model, emphasizing the\nneed to not only measure potential harms, but also understand how they arise by\ncharacterizing the data, mitigate harms once found, and balance any trade-offs.\nWe open-source our analysis code in hopes of encouraging broader measurements\nof bias in future LLMs.", + "authors": "David Esiobu, Xiaoqing Tan, Saghar Hosseini, Megan Ung, Yuchen Zhang, Jude Fernandes, Jane Dwivedi-Yu, Eleonora Presani, Adina Williams, Eric Michael Smith", + "published": "2023-11-29", + "updated": "2023-11-29", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2308.05345v3", + "title": "RTLLM: An Open-Source Benchmark for Design RTL Generation with Large Language Model", + "abstract": "Inspired by the recent success of large language models (LLMs) like ChatGPT,\nresearchers start to explore the adoption of LLMs for agile hardware design,\nsuch as generating design RTL based on natural-language instructions. However,\nin existing works, their target designs are all relatively simple and in a\nsmall scale, and proposed by the authors themselves, making a fair comparison\namong different LLM solutions challenging. In addition, many prior works only\nfocus on the design correctness, without evaluating the design qualities of\ngenerated design RTL. In this work, we propose an open-source benchmark named\nRTLLM, for generating design RTL with natural language instructions. To\nsystematically evaluate the auto-generated design RTL, we summarized three\nprogressive goals, named syntax goal, functionality goal, and design quality\ngoal. This benchmark can automatically provide a quantitative evaluation of any\ngiven LLM-based solution. Furthermore, we propose an easy-to-use yet\nsurprisingly effective prompt engineering technique named self-planning, which\nproves to significantly boost the performance of GPT-3.5 in our proposed\nbenchmark.", + "authors": "Yao Lu, Shang Liu, Qijun Zhang, Zhiyao Xie", + "published": "2023-08-10", + "updated": "2023-11-11", + "primary_cat": "cs.LG", + "cats": [ + "cs.LG", + "cs.AR" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2305.19118v1", + "title": "Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate", + "abstract": "Modern large language models (LLMs) like ChatGPT have shown remarkable\nperformance on general language tasks but still struggle on complex reasoning\ntasks, which drives the research on cognitive behaviors of LLMs to explore\nhuman-like problem-solving strategies. Along this direction, one representative\nstrategy is self-reflection, which asks an LLM to refine the solution with the\nfeedback generated by itself iteratively. However, our study shows that such\nreflection-style methods suffer from the Degeneration-of-Thought (DoT) problem:\nonce the LLM has established confidence in its solutions, it is unable to\ngenerate novel thoughts later through reflection even if its initial stance is\nincorrect. To address the DoT problem, we propose a Multi-Agent Debate (MAD)\nframework, in which multiple agents express their arguments in the state of\n\"tit for tat\" and a judge manages the debate process to obtain a final\nsolution. Clearly, our MAD framework encourages divergent thinking in LLMs\nwhich would be helpful for tasks that require deep levels of contemplation.\nExperiment results on two challenging datasets, commonsense machine translation\nand counter-intuitive arithmetic reasoning, demonstrate the effectiveness of\nour MAD framework. Extensive analyses suggest that the adaptive break of debate\nand the modest level of \"tit for tat\" state are required for MAD to obtain good\nperformance. Moreover, we find that LLMs might not be a fair judge if different\nLLMs are used for agents. Codes:\nhttps://github.com/Skytliang/Multi-Agents-Debate", + "authors": "Tian Liang, Zhiwei He, Wenxiang Jiao, Xing Wang, Yan Wang, Rui Wang, Yujiu Yang, Zhaopeng Tu, Shuming Shi", + "published": "2023-05-30", + "updated": "2023-05-30", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.05668v1", + "title": "CFaiRLLM: Consumer Fairness Evaluation in Large-Language Model Recommender System", + "abstract": "In the evolving landscape of recommender systems, the integration of Large\nLanguage Models (LLMs) such as ChatGPT marks a new era, introducing the concept\nof Recommendation via LLM (RecLLM). While these advancements promise\nunprecedented personalization and efficiency, they also bring to the fore\ncritical concerns regarding fairness, particularly in how recommendations might\ninadvertently perpetuate or amplify biases associated with sensitive user\nattributes. In order to address these concerns, our study introduces a\ncomprehensive evaluation framework, CFaiRLLM, aimed at evaluating (and thereby\nmitigating) biases on the consumer side within RecLLMs.\n Our research methodically assesses the fairness of RecLLMs by examining how\nrecommendations might vary with the inclusion of sensitive attributes such as\ngender, age, and their intersections, through both similarity alignment and\ntrue preference alignment. By analyzing recommendations generated under\ndifferent conditions-including the use of sensitive attributes in user\nprompts-our framework identifies potential biases in the recommendations\nprovided. A key part of our study involves exploring how different detailed\nstrategies for constructing user profiles (random, top-rated, recent) impact\nthe alignment between recommendations made without consideration of sensitive\nattributes and those that are sensitive-attribute-aware, highlighting the bias\nmechanisms within RecLLMs.\n The findings in our study highlight notable disparities in the fairness of\nrecommendations, particularly when sensitive attributes are integrated into the\nrecommendation process, either individually or in combination. The analysis\ndemonstrates that the choice of user profile sampling strategy plays a\nsignificant role in affecting fairness outcomes, highlighting the complexity of\nachieving fair recommendations in the era of LLMs.", + "authors": "Yashar Deldjoo, Tommaso di Noia", + "published": "2024-03-08", + "updated": "2024-03-08", + "primary_cat": "cs.IR", + "cats": [ + "cs.IR" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2308.10397v2", + "title": "FairMonitor: A Four-Stage Automatic Framework for Detecting Stereotypes and Biases in Large Language Models", + "abstract": "Detecting stereotypes and biases in Large Language Models (LLMs) can enhance\nfairness and reduce adverse impacts on individuals or groups when these LLMs\nare applied. However, the majority of existing methods focus on measuring the\nmodel's preference towards sentences containing biases and stereotypes within\ndatasets, which lacks interpretability and cannot detect implicit biases and\nstereotypes in the real world. To address this gap, this paper introduces a\nfour-stage framework to directly evaluate stereotypes and biases in the\ngenerated content of LLMs, including direct inquiry testing, serial or adapted\nstory testing, implicit association testing, and unknown situation testing.\nAdditionally, the paper proposes multi-dimensional evaluation metrics and\nexplainable zero-shot prompts for automated evaluation. Using the education\nsector as a case study, we constructed the Edu-FairMonitor based on the\nfour-stage framework, which encompasses 12,632 open-ended questions covering\nnine sensitive factors and 26 educational scenarios. Experimental results\nreveal varying degrees of stereotypes and biases in five LLMs evaluated on\nEdu-FairMonitor. Moreover, the results of our proposed automated evaluation\nmethod have shown a high correlation with human annotations.", + "authors": "Yanhong Bai, Jiabao Zhao, Jinxin Shi, Tingjiang Wei, Xingjiao Wu, Liang He", + "published": "2023-08-21", + "updated": "2023-10-27", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2401.00588v1", + "title": "Fairness in Serving Large Language Models", + "abstract": "High-demand LLM inference services (e.g., ChatGPT and BARD) support a wide\nrange of requests from short chat conversations to long document reading. To\nensure that all client requests are processed fairly, most major LLM inference\nservices have request rate limits, to ensure that no client can dominate the\nrequest queue. However, this rudimentary notion of fairness also results in\nunder-utilization of the resources and poor client experience when there is\nspare capacity. While there is a rich literature on fair scheduling, serving\nLLMs presents new challenges due to their unpredictable request lengths and\ntheir unique batching characteristics on parallel accelerators. This paper\nintroduces the definition of LLM serving fairness based on a cost function that\naccounts for the number of input and output tokens processed. To achieve\nfairness in serving, we propose a novel scheduling algorithm, the Virtual Token\nCounter (VTC), a fair scheduler based on the continuous batching mechanism. We\nprove a 2x tight upper bound on the service difference between two backlogged\nclients, adhering to the requirement of work-conserving. Through extensive\nexperiments, we demonstrate the superior performance of VTC in ensuring\nfairness, especially in contrast to other baseline methods, which exhibit\nshortcomings under various conditions.", + "authors": "Ying Sheng, Shiyi Cao, Dacheng Li, Banghua Zhu, Zhuohan Li, Danyang Zhuo, Joseph E. Gonzalez, Ion Stoica", + "published": "2023-12-31", + "updated": "2023-12-31", + "primary_cat": "cs.AI", + "cats": [ + "cs.AI", + "cs.LG", + "cs.PF" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.11764v1", + "title": "ChatGPT Based Data Augmentation for Improved Parameter-Efficient Debiasing of LLMs", + "abstract": "Large Language models (LLMs), while powerful, exhibit harmful social biases.\nDebiasing is often challenging due to computational costs, data constraints,\nand potential degradation of multi-task language capabilities. This work\nintroduces a novel approach utilizing ChatGPT to generate synthetic training\ndata, aiming to enhance the debiasing of LLMs. We propose two strategies:\nTargeted Prompting, which provides effective debiasing for known biases but\nnecessitates prior specification of bias in question; and General Prompting,\nwhich, while slightly less effective, offers debiasing across various\ncategories. We leverage resource-efficient LLM debiasing using adapter tuning\nand compare the effectiveness of our synthetic data to existing debiasing\ndatasets. Our results reveal that: (1) ChatGPT can efficiently produce\nhigh-quality training data for debiasing other LLMs; (2) data produced via our\napproach surpasses existing datasets in debiasing performance while also\npreserving internal knowledge of a pre-trained LLM; and (3) synthetic data\nexhibits generalizability across categories, effectively mitigating various\nbiases, including intersectional ones. These findings underscore the potential\nof synthetic data in advancing the fairness of LLMs with minimal retraining\ncost.", + "authors": "Pengrui Han, Rafal Kocielnik, Adhithya Saravanan, Roy Jiang, Or Sharir, Anima Anandkumar", + "published": "2024-02-19", + "updated": "2024-02-19", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.CY", + "68T50", + "I.2.7; K.4.1" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2401.01262v2", + "title": "Fairness Certification for Natural Language Processing and Large Language Models", + "abstract": "Natural Language Processing (NLP) plays an important role in our daily lives,\nparticularly due to the enormous progress of Large Language Models (LLM).\nHowever, NLP has many fairness-critical use cases, e.g., as an expert system in\nrecruitment or as an LLM-based tutor in education. Since NLP is based on human\nlanguage, potentially harmful biases can diffuse into NLP systems and produce\nunfair results, discriminate against minorities or generate legal issues.\nHence, it is important to develop a fairness certification for NLP approaches.\nWe follow a qualitative research approach towards a fairness certification for\nNLP. In particular, we have reviewed a large body of literature on algorithmic\nfairness, and we have conducted semi-structured expert interviews with a wide\nrange of experts from that area. We have systematically devised six fairness\ncriteria for NLP, which can be further refined into 18 sub-categories. Our\ncriteria offer a foundation for operationalizing and testing processes to\ncertify fairness, both from the perspective of the auditor and the audited\norganization.", + "authors": "Vincent Freiberger, Erik Buchmann", + "published": "2024-01-02", + "updated": "2024-01-03", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.CY", + "cs.LG", + "68T50", + "I.2.7" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.14607v2", + "title": "Confronting LLMs with Traditional ML: Rethinking the Fairness of Large Language Models in Tabular Classifications", + "abstract": "Recent literature has suggested the potential of using large language models\n(LLMs) to make classifications for tabular tasks. However, LLMs have been shown\nto exhibit harmful social biases that reflect the stereotypes and inequalities\npresent in society. To this end, as well as the widespread use of tabular data\nin many high-stake applications, it is important to explore the following\nquestions: what sources of information do LLMs draw upon when making\nclassifications for tabular tasks; whether and to what extent are LLM\nclassifications for tabular data influenced by social biases and stereotypes;\nand what are the consequential implications for fairness?\n Through a series of experiments, we delve into these questions and show that\nLLMs tend to inherit social biases from their training data which significantly\nimpact their fairness in tabular classification tasks. Furthermore, our\ninvestigations show that in the context of bias mitigation, though in-context\nlearning and finetuning have a moderate effect, the fairness metric gap between\ndifferent subgroups is still larger than that in traditional machine learning\nmodels, such as Random Forest and shallow Neural Networks. This observation\nemphasizes that the social biases are inherent within the LLMs themselves and\ninherited from their pretraining corpus, not only from the downstream task\ndatasets. Besides, we demonstrate that label-flipping of in-context examples\ncan significantly reduce biases, further highlighting the presence of inherent\nbias within LLMs.", + "authors": "Yanchen Liu, Srishti Gautam, Jiaqi Ma, Himabindu Lakkaraju", + "published": "2023-10-23", + "updated": "2024-04-02", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.13095v1", + "title": "Enhancing Logical Reasoning in Large Language Models to Facilitate Legal Applications", + "abstract": "Language serves as a vehicle for conveying thought, enabling communication\namong individuals. The ability to distinguish between diverse concepts,\nidentify fairness and injustice, and comprehend a range of legal notions\nfundamentally relies on logical reasoning. Large Language Models (LLMs) attempt\nto emulate human language understanding and generation, but their competency in\nlogical reasoning remains limited. This paper seeks to address the\nphilosophical question: How can we effectively teach logical reasoning to LLMs\nwhile maintaining a deep understanding of the intricate relationship between\nlanguage and logic? By focusing on bolstering LLMs' capabilities in logical\nreasoning, we aim to expand their applicability in law and other\nlogic-intensive disciplines. To this end, we propose a Reinforcement Learning\nfrom Logical Feedback (RLLF) approach, which serves as a potential framework\nfor refining LLMs' reasoning capacities. Through RLLF and a revised evaluation\nmethodology, we explore new avenues for research in this domain and contribute\nto the development of LLMs capable of handling complex legal reasoning tasks\nwhile acknowledging the fundamental connection between language and logic.", + "authors": "Ha-Thanh Nguyen, Wachara Fungwacharakorn, Ken Satoh", + "published": "2023-11-22", + "updated": "2023-11-22", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.04892v2", + "title": "Bias Runs Deep: Implicit Reasoning Biases in Persona-Assigned LLMs", + "abstract": "Recent works have showcased the ability of LLMs to embody diverse personas in\ntheir responses, exemplified by prompts like 'You are Yoda. Explain the Theory\nof Relativity.' While this ability allows personalization of LLMs and enables\nhuman behavior simulation, its effect on LLMs' capabilities remains unclear. To\nfill this gap, we present the first extensive study of the unintended\nside-effects of persona assignment on the ability of LLMs to perform basic\nreasoning tasks. Our study covers 24 reasoning datasets, 4 LLMs, and 19 diverse\npersonas (e.g. an Asian person) spanning 5 socio-demographic groups. Our\nexperiments unveil that LLMs harbor deep rooted bias against various\nsocio-demographics underneath a veneer of fairness. While they overtly reject\nstereotypes when explicitly asked ('Are Black people less skilled at\nmathematics?'), they manifest stereotypical and erroneous presumptions when\nasked to answer questions while adopting a persona. These can be observed as\nabstentions in responses, e.g., 'As a Black person, I can't answer this\nquestion as it requires math knowledge', and generally result in a substantial\nperformance drop. Our experiments with ChatGPT-3.5 show that this bias is\nubiquitous - 80% of our personas demonstrate bias; it is significant - some\ndatasets show performance drops of 70%+; and can be especially harmful for\ncertain groups - some personas suffer statistically significant drops on 80%+\nof the datasets. Overall, all 4 LLMs exhibit this bias to varying extents, with\nGPT-4-Turbo showing the least but still a problematic amount of bias (evident\nin 42% of the personas). Further analysis shows that these persona-induced\nerrors can be hard-to-discern and hard-to-avoid. Our findings serve as a\ncautionary tale that the practice of assigning personas to LLMs - a trend on\nthe rise - can surface their deep-rooted biases and have unforeseeable and\ndetrimental side-effects.", + "authors": "Shashank Gupta, Vaishnavi Shrivastava, Ameet Deshpande, Ashwin Kalyan, Peter Clark, Ashish Sabharwal, Tushar Khot", + "published": "2023-11-08", + "updated": "2024-01-27", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.13840v1", + "title": "Whose Side Are You On? Investigating the Political Stance of Large Language Models", + "abstract": "Large Language Models (LLMs) have gained significant popularity for their\napplication in various everyday tasks such as text generation, summarization,\nand information retrieval. As the widespread adoption of LLMs continues to\nsurge, it becomes increasingly crucial to ensure that these models yield\nresponses that are politically impartial, with the aim of preventing\ninformation bubbles, upholding fairness in representation, and mitigating\nconfirmation bias. In this paper, we propose a quantitative framework and\npipeline designed to systematically investigate the political orientation of\nLLMs. Our investigation delves into the political alignment of LLMs across a\nspectrum of eight polarizing topics, spanning from abortion to LGBTQ issues.\nAcross topics, the results indicate that LLMs exhibit a tendency to provide\nresponses that closely align with liberal or left-leaning perspectives rather\nthan conservative or right-leaning ones when user queries include details\npertaining to occupation, race, or political affiliation. The findings\npresented in this study not only reaffirm earlier observations regarding the\nleft-leaning characteristics of LLMs but also surface particular attributes,\nsuch as occupation, that are particularly susceptible to such inclinations even\nwhen directly steered towards conservatism. As a recommendation to avoid these\nmodels providing politicised responses, users should be mindful when crafting\nqueries, and exercise caution in selecting neutral prompt language.", + "authors": "Pagnarasmey Pit, Xingjun Ma, Mike Conway, Qingyu Chen, James Bailey, Henry Pit, Putrasmey Keo, Watey Diep, Yu-Gang Jiang", + "published": "2024-03-15", + "updated": "2024-03-15", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.SI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.07884v2", + "title": "Fair Abstractive Summarization of Diverse Perspectives", + "abstract": "People from different social and demographic groups express diverse\nperspectives and conflicting opinions on a broad set of topics such as product\nreviews, healthcare, law, and politics. A fair summary should provide a\ncomprehensive coverage of diverse perspectives without underrepresenting\ncertain groups. However, current work in summarization metrics and Large\nLanguage Models (LLMs) evaluation has not explored fair abstractive\nsummarization. In this paper, we systematically investigate fair abstractive\nsummarization for user-generated data. We first formally define fairness in\nabstractive summarization as not underrepresenting perspectives of any groups\nof people, and we propose four reference-free automatic metrics by measuring\nthe differences between target and source perspectives. We evaluate nine LLMs,\nincluding three GPT models, four LLaMA models, PaLM 2, and Claude, on six\ndatasets collected from social media, online reviews, and recorded transcripts.\nExperiments show that both the model-generated and the human-written reference\nsummaries suffer from low fairness. We conduct a comprehensive analysis of the\ncommon factors influencing fairness and propose three simple but effective\nmethods to alleviate unfair summarization. Our dataset and code are available\nat https://github.com/psunlpgroup/FairSumm.", + "authors": "Yusen Zhang, Nan Zhang, Yixin Liu, Alexander Fabbri, Junru Liu, Ryo Kamoi, Xiaoxin Lu, Caiming Xiong, Jieyu Zhao, Dragomir Radev, Kathleen McKeown, Rui Zhang", + "published": "2023-11-14", + "updated": "2024-03-30", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.12150v1", + "title": "Your Large Language Model is Secretly a Fairness Proponent and You Should Prompt it Like One", + "abstract": "The widespread adoption of large language models (LLMs) underscores the\nurgent need to ensure their fairness. However, LLMs frequently present dominant\nviewpoints while ignoring alternative perspectives from minority parties,\nresulting in potential biases. We hypothesize that these fairness-violating\nbehaviors occur because LLMs express their viewpoints using a human personality\nthat represents the majority of training data. In response to this, we validate\nthat prompting LLMs with specific roles can allow LLMs to express diverse\nviewpoints. Building on this insight and observation, we develop FairThinking,\na pipeline designed to automatically generate roles that enable LLMs to\narticulate diverse perspectives for fair expressions. To evaluate FairThinking,\nwe create a dataset with a thousand items covering three fairness-related\ntopics and conduct experiments on GPT-3.5, GPT-4, Llama2, and Mistral to\ndemonstrate its superior performance.", + "authors": "Tianlin Li, Xiaoyu Zhang, Chao Du, Tianyu Pang, Qian Liu, Qing Guo, Chao Shen, Yang Liu", + "published": "2024-02-19", + "updated": "2024-02-19", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "I.2; J.4" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.08472v1", + "title": "Selecting Shots for Demographic Fairness in Few-Shot Learning with Large Language Models", + "abstract": "Recently, work in NLP has shifted to few-shot (in-context) learning, with\nlarge language models (LLMs) performing well across a range of tasks. However,\nwhile fairness evaluations have become a standard for supervised methods,\nlittle is known about the fairness of LLMs as prediction systems. Further,\ncommon standard methods for fairness involve access to models weights or are\napplied during finetuning, which are not applicable in few-shot learning. Do\nLLMs exhibit prediction biases when used for standard NLP tasks? In this work,\nwe explore the effect of shots, which directly affect the performance of\nmodels, on the fairness of LLMs as NLP classification systems. We consider how\ndifferent shot selection strategies, both existing and new demographically\nsensitive methods, affect model fairness across three standard fairness\ndatasets. We discuss how future work can include LLM fairness evaluations.", + "authors": "Carlos Aguirre, Kuleen Sasse, Isabel Cachola, Mark Dredze", + "published": "2023-11-14", + "updated": "2023-11-14", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.15007v1", + "title": "Did the Neurons Read your Book? Document-level Membership Inference for Large Language Models", + "abstract": "With large language models (LLMs) poised to become embedded in our daily\nlives, questions are starting to be raised about the dataset(s) they learned\nfrom. These questions range from potential bias or misinformation LLMs could\nretain from their training data to questions of copyright and fair use of\nhuman-generated text. However, while these questions emerge, developers of the\nrecent state-of-the-art LLMs become increasingly reluctant to disclose details\non their training corpus. We here introduce the task of document-level\nmembership inference for real-world LLMs, i.e. inferring whether the LLM has\nseen a given document during training or not. First, we propose a procedure for\nthe development and evaluation of document-level membership inference for LLMs\nby leveraging commonly used data sources for training and the model release\ndate. We then propose a practical, black-box method to predict document-level\nmembership and instantiate it on OpenLLaMA-7B with both books and academic\npapers. We show our methodology to perform very well, reaching an impressive\nAUC of 0.856 for books and 0.678 for papers. We then show our approach to\noutperform the sentence-level membership inference attacks used in the privacy\nliterature for the document-level membership task. We finally evaluate whether\nsmaller models might be less sensitive to document-level inference and show\nOpenLLaMA-3B to be approximately as sensitive as OpenLLaMA-7B to our approach.\nTaken together, our results show that accurate document-level membership can be\ninferred for LLMs, increasing the transparency of technology poised to change\nour lives.", + "authors": "Matthieu Meeus, Shubham Jain, Marek Rei, Yves-Alexandre de Montjoye", + "published": "2023-10-23", + "updated": "2023-10-23", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.CR", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.11406v2", + "title": "Don't Go To Extremes: Revealing the Excessive Sensitivity and Calibration Limitations of LLMs in Implicit Hate Speech Detection", + "abstract": "The fairness and trustworthiness of Large Language Models (LLMs) are\nreceiving increasing attention. Implicit hate speech, which employs indirect\nlanguage to convey hateful intentions, occupies a significant portion of\npractice. However, the extent to which LLMs effectively address this issue\nremains insufficiently examined. This paper delves into the capability of LLMs\nto detect implicit hate speech (Classification Task) and express confidence in\ntheir responses (Calibration Task). Our evaluation meticulously considers\nvarious prompt patterns and mainstream uncertainty estimation methods. Our\nfindings highlight that LLMs exhibit two extremes: (1) LLMs display excessive\nsensitivity towards groups or topics that may cause fairness issues, resulting\nin misclassifying benign statements as hate speech. (2) LLMs' confidence scores\nfor each method excessively concentrate on a fixed range, remaining unchanged\nregardless of the dataset's complexity. Consequently, the calibration\nperformance is heavily reliant on primary classification accuracy. These\ndiscoveries unveil new limitations of LLMs, underscoring the need for caution\nwhen optimizing models to ensure they do not veer towards extremes. This serves\nas a reminder to carefully consider sensitivity and confidence in the pursuit\nof model fairness.", + "authors": "Min Zhang, Jianfeng He, Taoran Ji, Chang-Tien Lu", + "published": "2024-02-18", + "updated": "2024-02-26", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2312.14769v3", + "title": "Large Language Model (LLM) Bias Index -- LLMBI", + "abstract": "The Large Language Model Bias Index (LLMBI) is a pioneering approach designed\nto quantify and address biases inherent in large language models (LLMs), such\nas GPT-4. We recognise the increasing prevalence and impact of LLMs across\ndiverse sectors. This research introduces a novel metric, LLMBI, to\nsystematically measure and mitigate biases potentially skewing model responses.\nWe formulated LLMBI using a composite scoring system incorporating multiple\ndimensions of bias, including but not limited to age, gender, and racial\nbiases. To operationalise this metric, we engaged in a multi-step process\ninvolving collecting and annotating LLM responses, applying sophisticated\nNatural Language Processing (NLP) techniques for bias detection, and computing\nthe LLMBI score through a specially crafted mathematical formula. The formula\nintegrates weighted averages of various bias dimensions, a penalty for dataset\ndiversity deficiencies, and a correction for sentiment biases. Our empirical\nanalysis, conducted using responses from OpenAI's API, employs advanced\nsentiment analysis as a representative method for bias detection. The research\nreveals LLMs, whilst demonstrating impressive capabilities in text generation,\nexhibit varying degrees of bias across different dimensions. LLMBI provides a\nquantifiable measure to compare biases across models and over time, offering a\nvital tool for systems engineers, researchers and regulators in enhancing the\nfairness and reliability of LLMs. It highlights the potential of LLMs in\nmimicking unbiased human-like responses. Additionally, it underscores the\nnecessity of continuously monitoring and recalibrating such models to align\nwith evolving societal norms and ethical standards.", + "authors": "Abiodun Finbarrs Oketunji, Muhammad Anas, Deepthi Saina", + "published": "2023-12-22", + "updated": "2023-12-29", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.CY", + "cs.LG", + "I.2.7" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2309.09397v1", + "title": "Do Large GPT Models Discover Moral Dimensions in Language Representations? A Topological Study Of Sentence Embeddings", + "abstract": "As Large Language Models are deployed within Artificial Intelligence systems,\nthat are increasingly integrated with human society, it becomes more important\nthan ever to study their internal structures. Higher level abilities of LLMs\nsuch as GPT-3.5 emerge in large part due to informative language\nrepresentations they induce from raw text data during pre-training on trillions\nof words. These embeddings exist in vector spaces of several thousand\ndimensions, and their processing involves mapping between multiple vector\nspaces, with total number of parameters on the order of trillions. Furthermore,\nthese language representations are induced by gradient optimization, resulting\nin a black box system that is hard to interpret. In this paper, we take a look\nat the topological structure of neuronal activity in the \"brain\" of Chat-GPT's\nfoundation language model, and analyze it with respect to a metric representing\nthe notion of fairness. We develop a novel approach to visualize GPT's moral\ndimensions. We first compute a fairness metric, inspired by social psychology\nliterature, to identify factors that typically influence fairness assessments\nin humans, such as legitimacy, need, and responsibility. Subsequently, we\nsummarize the manifold's shape using a lower-dimensional simplicial complex,\nwhose topology is derived from this metric. We color it with a heat map\nassociated with this fairness metric, producing human-readable visualizations\nof the high-dimensional sentence manifold. Our results show that sentence\nembeddings based on GPT-3.5 can be decomposed into two submanifolds\ncorresponding to fair and unfair moral judgments. This indicates that GPT-based\nlanguage models develop a moral dimension within their representation spaces\nand induce an understanding of fairness during their training process.", + "authors": "Stephen Fitz", + "published": "2023-09-17", + "updated": "2023-09-17", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.CY", + "cs.LG", + "cs.NE" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2307.03838v2", + "title": "RADAR: Robust AI-Text Detection via Adversarial Learning", + "abstract": "Recent advances in large language models (LLMs) and the intensifying\npopularity of ChatGPT-like applications have blurred the boundary of\nhigh-quality text generation between humans and machines. However, in addition\nto the anticipated revolutionary changes to our technology and society, the\ndifficulty of distinguishing LLM-generated texts (AI-text) from human-generated\ntexts poses new challenges of misuse and fairness, such as fake content\ngeneration, plagiarism, and false accusations of innocent writers. While\nexisting works show that current AI-text detectors are not robust to LLM-based\nparaphrasing, this paper aims to bridge this gap by proposing a new framework\ncalled RADAR, which jointly trains a robust AI-text detector via adversarial\nlearning. RADAR is based on adversarial training of a paraphraser and a\ndetector. The paraphraser's goal is to generate realistic content to evade\nAI-text detection. RADAR uses the feedback from the detector to update the\nparaphraser, and vice versa. Evaluated with 8 different LLMs (Pythia, Dolly\n2.0, Palmyra, Camel, GPT-J, Dolly 1.0, LLaMA, and Vicuna) across 4 datasets,\nexperimental results show that RADAR significantly outperforms existing AI-text\ndetection methods, especially when paraphrasing is in place. We also identify\nthe strong transferability of RADAR from instruction-tuned LLMs to other LLMs,\nand evaluate the improved capability of RADAR via GPT-3.5-Turbo.", + "authors": "Xiaomeng Hu, Pin-Yu Chen, Tsung-Yi Ho", + "published": "2023-07-07", + "updated": "2023-10-24", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2401.00625v2", + "title": "Beyond Efficiency: A Systematic Survey of Resource-Efficient Large Language Models", + "abstract": "The burgeoning field of Large Language Models (LLMs), exemplified by\nsophisticated models like OpenAI's ChatGPT, represents a significant\nadvancement in artificial intelligence. These models, however, bring forth\nsubstantial challenges in the high consumption of computational, memory,\nenergy, and financial resources, especially in environments with limited\nresource capabilities. This survey aims to systematically address these\nchallenges by reviewing a broad spectrum of techniques designed to enhance the\nresource efficiency of LLMs. We categorize methods based on their optimization\nfocus: computational, memory, energy, financial, and network resources and\ntheir applicability across various stages of an LLM's lifecycle, including\narchitecture design, pretraining, finetuning, and system design. Additionally,\nthe survey introduces a nuanced categorization of resource efficiency\ntechniques by their specific resource types, which uncovers the intricate\nrelationships and mappings between various resources and corresponding\noptimization techniques. A standardized set of evaluation metrics and datasets\nis also presented to facilitate consistent and fair comparisons across\ndifferent models and techniques. By offering a comprehensive overview of the\ncurrent sota and identifying open research avenues, this survey serves as a\nfoundational reference for researchers and practitioners, aiding them in\ndeveloping more sustainable and efficient LLMs in a rapidly evolving landscape.", + "authors": "Guangji Bai, Zheng Chai, Chen Ling, Shiyu Wang, Jiaying Lu, Nan Zhang, Tingwei Shi, Ziyang Yu, Mengdan Zhu, Yifei Zhang, Carl Yang, Yue Cheng, Liang Zhao", + "published": "2024-01-01", + "updated": "2024-01-04", + "primary_cat": "cs.LG", + "cats": [ + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2305.12090v1", + "title": "UP5: Unbiased Foundation Model for Fairness-aware Recommendation", + "abstract": "Recent advancements in foundation models such as large language models (LLM)\nhave propelled them to the forefront of recommender systems (RS). Moreover,\nfairness in RS is critical since many users apply it for decision-making and\ndemand fulfillment. However, at present, there is a lack of understanding\nregarding the level of fairness exhibited by recommendation foundation models\nand the appropriate methods for equitably treating different groups of users in\nfoundation models. In this paper, we focus on user-side unfairness problem and\nshow through a thorough examination that there is unfairness involved in LLMs\nthat lead to unfair recommendation results. To eliminate bias from LLM for\nfairness-aware recommendation, we introduce a novel Unbiased P5 (UP5)\nfoundation model based on Counterfactually-Fair-Prompting (CFP) techniques. CFP\nincludes two sub-modules: a personalized prefix prompt that enhances fairness\nwith respect to individual sensitive attributes, and a Prompt Mixture that\nintegrates multiple counterfactually-fair prompts for a set of sensitive\nattributes. Experiments are conducted on two real-world datasets, MovieLens-1M\nand Insurance, and results are compared with both matching-based and\nsequential-based fairness-aware recommendation models. The results show that\nUP5 achieves better recommendation performance and meanwhile exhibits a high\nlevel of fairness.", + "authors": "Wenyue Hua, Yingqiang Ge, Shuyuan Xu, Jianchao Ji, Yongfeng Zhang", + "published": "2023-05-20", + "updated": "2023-05-20", + "primary_cat": "cs.IR", + "cats": [ + "cs.IR", + "cs.AI", + "cs.CL", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2308.10149v2", + "title": "A Survey on Fairness in Large Language Models", + "abstract": "Large Language Models (LLMs) have shown powerful performance and development\nprospects and are widely deployed in the real world. However, LLMs can capture\nsocial biases from unprocessed training data and propagate the biases to\ndownstream tasks. Unfair LLM systems have undesirable social impacts and\npotential harms. In this paper, we provide a comprehensive review of related\nresearch on fairness in LLMs. Considering the influence of parameter magnitude\nand training paradigm on research strategy, we divide existing fairness\nresearch into oriented to medium-sized LLMs under pre-training and fine-tuning\nparadigms and oriented to large-sized LLMs under prompting paradigms. First,\nfor medium-sized LLMs, we introduce evaluation metrics and debiasing methods\nfrom the perspectives of intrinsic bias and extrinsic bias, respectively. Then,\nfor large-sized LLMs, we introduce recent fairness research, including fairness\nevaluation, reasons for bias, and debiasing methods. Finally, we discuss and\nprovide insight on the challenges and future directions for the development of\nfairness in LLMs.", + "authors": "Yingji Li, Mengnan Du, Rui Song, Xin Wang, Ying Wang", + "published": "2023-08-20", + "updated": "2024-02-21", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2305.01937v1", + "title": "Can Large Language Models Be an Alternative to Human Evaluations?", + "abstract": "Human evaluation is indispensable and inevitable for assessing the quality of\ntexts generated by machine learning models or written by humans. However, human\nevaluation is very difficult to reproduce and its quality is notoriously\nunstable, hindering fair comparisons among different natural language\nprocessing (NLP) models and algorithms. Recently, large language models (LLMs)\nhave demonstrated exceptional performance on unseen tasks when only the task\ninstructions are provided. In this paper, we explore if such an ability of the\nLLMs can be used as an alternative to human evaluation. We present the LLMs\nwith the exact same instructions, samples to be evaluated, and questions used\nto conduct human evaluation, and then ask the LLMs to generate responses to\nthose questions; we dub this LLM evaluation. We use human evaluation and LLM\nevaluation to evaluate the texts in two NLP tasks: open-ended story generation\nand adversarial attacks. We show that the result of LLM evaluation is\nconsistent with the results obtained by expert human evaluation: the texts\nrated higher by human experts are also rated higher by the LLMs. We also find\nthat the results of LLM evaluation are stable over different formatting of the\ntask instructions and the sampling algorithm used to generate the answer. We\nare the first to show the potential of using LLMs to assess the quality of\ntexts and discuss the limitations and ethical considerations of LLM evaluation.", + "authors": "Cheng-Han Chiang, Hung-yi Lee", + "published": "2023-05-03", + "updated": "2023-05-03", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.HC" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2401.08495v2", + "title": "Large Language Models Portray Socially Subordinate Groups as More Homogeneous, Consistent with a Bias Observed in Humans", + "abstract": "Large language models (LLMs) are becoming pervasive in everyday life, yet\ntheir propensity to reproduce biases inherited from training data remains a\npressing concern. Prior investigations into bias in LLMs have focused on the\nassociation of social groups with stereotypical attributes. However, this is\nonly one form of human bias such systems may reproduce. We investigate a new\nform of bias in LLMs that resembles a social psychological phenomenon where\nsocially subordinate groups are perceived as more homogeneous than socially\ndominant groups. We had ChatGPT, a state-of-the-art LLM, generate texts about\nintersectional group identities and compared those texts on measures of\nhomogeneity. We consistently found that ChatGPT portrayed African, Asian, and\nHispanic Americans as more homogeneous than White Americans, indicating that\nthe model described racial minority groups with a narrower range of human\nexperience. ChatGPT also portrayed women as more homogeneous than men, but\nthese differences were small. Finally, we found that the effect of gender\ndiffered across racial/ethnic groups such that the effect of gender was\nconsistent within African and Hispanic Americans but not within Asian and White\nAmericans. We argue that the tendency of LLMs to describe groups as less\ndiverse risks perpetuating stereotypes and discriminatory behavior.", + "authors": "Messi H. J. Lee, Jacob M. Montgomery, Calvin K. Lai", + "published": "2024-01-16", + "updated": "2024-04-26", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.05694v1", + "title": "A Survey of Large Language Models for Healthcare: from Data, Technology, and Applications to Accountability and Ethics", + "abstract": "The utilization of large language models (LLMs) in the Healthcare domain has\ngenerated both excitement and concern due to their ability to effectively\nrespond to freetext queries with certain professional knowledge. This survey\noutlines the capabilities of the currently developed LLMs for Healthcare and\nexplicates their development process, with the aim of providing an overview of\nthe development roadmap from traditional Pretrained Language Models (PLMs) to\nLLMs. Specifically, we first explore the potential of LLMs to enhance the\nefficiency and effectiveness of various Healthcare applications highlighting\nboth the strengths and limitations. Secondly, we conduct a comparison between\nthe previous PLMs and the latest LLMs, as well as comparing various LLMs with\neach other. Then we summarize related Healthcare training data, training\nmethods, optimization strategies, and usage. Finally, the unique concerns\nassociated with deploying LLMs in Healthcare settings are investigated,\nparticularly regarding fairness, accountability, transparency and ethics. Our\nsurvey provide a comprehensive investigation from perspectives of both computer\nscience and Healthcare specialty. Besides the discussion about Healthcare\nconcerns, we supports the computer science community by compiling a collection\nof open source resources, such as accessible datasets, the latest\nmethodologies, code implementations, and evaluation benchmarks in the Github.\nSummarily, we contend that a significant paradigm shift is underway,\ntransitioning from PLMs to LLMs. This shift encompasses a move from\ndiscriminative AI approaches to generative AI approaches, as well as a shift\nfrom model-centered methodologies to datacentered methodologies.", + "authors": "Kai He, Rui Mao, Qika Lin, Yucheng Ruan, Xiang Lan, Mengling Feng, Erik Cambria", + "published": "2023-10-09", + "updated": "2023-10-09", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.04489v1", + "title": "De-amplifying Bias from Differential Privacy in Language Model Fine-tuning", + "abstract": "Fairness and privacy are two important values machine learning (ML)\npractitioners often seek to operationalize in models. Fairness aims to reduce\nmodel bias for social/demographic sub-groups. Privacy via differential privacy\n(DP) mechanisms, on the other hand, limits the impact of any individual's\ntraining data on the resulting model. The trade-offs between privacy and\nfairness goals of trustworthy ML pose a challenge to those wishing to address\nboth. We show that DP amplifies gender, racial, and religious bias when\nfine-tuning large language models (LLMs), producing models more biased than\nones fine-tuned without DP. We find the cause of the amplification to be a\ndisparity in convergence of gradients across sub-groups. Through the case of\nbinary gender bias, we demonstrate that Counterfactual Data Augmentation (CDA),\na known method for addressing bias, also mitigates bias amplification by DP. As\na consequence, DP and CDA together can be used to fine-tune models while\nmaintaining both fairness and privacy.", + "authors": "Sanjari Srivastava, Piotr Mardziel, Zhikhun Zhang, Archana Ahlawat, Anupam Datta, John C Mitchell", + "published": "2024-02-07", + "updated": "2024-02-07", + "primary_cat": "cs.LG", + "cats": [ + "cs.LG", + "cs.CR", + "cs.CY", + "stat.ME" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2405.01769v1", + "title": "A Survey on Large Language Models for Critical Societal Domains: Finance, Healthcare, and Law", + "abstract": "In the fast-evolving domain of artificial intelligence, large language models\n(LLMs) such as GPT-3 and GPT-4 are revolutionizing the landscapes of finance,\nhealthcare, and law: domains characterized by their reliance on professional\nexpertise, challenging data acquisition, high-stakes, and stringent regulatory\ncompliance. This survey offers a detailed exploration of the methodologies,\napplications, challenges, and forward-looking opportunities of LLMs within\nthese high-stakes sectors. We highlight the instrumental role of LLMs in\nenhancing diagnostic and treatment methodologies in healthcare, innovating\nfinancial analytics, and refining legal interpretation and compliance\nstrategies. Moreover, we critically examine the ethics for LLM applications in\nthese fields, pointing out the existing ethical concerns and the need for\ntransparent, fair, and robust AI systems that respect regulatory norms. By\npresenting a thorough review of current literature and practical applications,\nwe showcase the transformative impact of LLMs, and outline the imperative for\ninterdisciplinary cooperation, methodological advancements, and ethical\nvigilance. Through this lens, we aim to spark dialogue and inspire future\nresearch dedicated to maximizing the benefits of LLMs while mitigating their\nrisks in these precision-dependent sectors. To facilitate future research on\nLLMs in these critical societal domains, we also initiate a reading list that\ntracks the latest advancements under this topic, which will be continually\nupdated: \\url{https://github.com/czyssrs/LLM_X_papers}.", + "authors": "Zhiyu Zoey Chen, Jing Ma, Xinlu Zhang, Nan Hao, An Yan, Armineh Nourbakhsh, Xianjun Yang, Julian McAuley, Linda Petzold, William Yang Wang", + "published": "2024-05-02", + "updated": "2024-05-02", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.18502v1", + "title": "Few-Shot Fairness: Unveiling LLM's Potential for Fairness-Aware Classification", + "abstract": "Employing Large Language Models (LLM) in various downstream applications such\nas classification is crucial, especially for smaller companies lacking the\nexpertise and resources required for fine-tuning a model. Fairness in LLMs\nhelps ensure inclusivity, equal representation based on factors such as race,\ngender and promotes responsible AI deployment. As the use of LLMs has become\nincreasingly prevalent, it is essential to assess whether LLMs can generate\nfair outcomes when subjected to considerations of fairness. In this study, we\nintroduce a framework outlining fairness regulations aligned with various\nfairness definitions, with each definition being modulated by varying degrees\nof abstraction. We explore the configuration for in-context learning and the\nprocedure for selecting in-context demonstrations using RAG, while\nincorporating fairness rules into the process. Experiments conducted with\ndifferent LLMs indicate that GPT-4 delivers superior results in terms of both\naccuracy and fairness compared to other models. This work is one of the early\nattempts to achieve fairness in prediction tasks by utilizing LLMs through\nin-context learning.", + "authors": "Garima Chhikara, Anurag Sharma, Kripabandhu Ghosh, Abhijnan Chakraborty", + "published": "2024-02-28", + "updated": "2024-02-28", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2305.03514v3", + "title": "Can Large Language Models Transform Computational Social Science?", + "abstract": "Large Language Models (LLMs) are capable of successfully performing many\nlanguage processing tasks zero-shot (without training data). If zero-shot LLMs\ncan also reliably classify and explain social phenomena like persuasiveness and\npolitical ideology, then LLMs could augment the Computational Social Science\n(CSS) pipeline in important ways. This work provides a road map for using LLMs\nas CSS tools. Towards this end, we contribute a set of prompting best practices\nand an extensive evaluation pipeline to measure the zero-shot performance of 13\nlanguage models on 25 representative English CSS benchmarks. On taxonomic\nlabeling tasks (classification), LLMs fail to outperform the best fine-tuned\nmodels but still achieve fair levels of agreement with humans. On free-form\ncoding tasks (generation), LLMs produce explanations that often exceed the\nquality of crowdworkers' gold references. We conclude that the performance of\ntoday's LLMs can augment the CSS research pipeline in two ways: (1) serving as\nzero-shot data annotators on human annotation teams, and (2) bootstrapping\nchallenging creative generation tasks (e.g., explaining the underlying\nattributes of a text). In summary, LLMs are posed to meaningfully participate\nin social science analysis in partnership with humans.", + "authors": "Caleb Ziems, William Held, Omar Shaikh, Jiaao Chen, Zhehao Zhang, Diyi Yang", + "published": "2023-04-12", + "updated": "2024-02-26", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2312.15398v1", + "title": "Fairness-Aware Structured Pruning in Transformers", + "abstract": "The increasing size of large language models (LLMs) has introduced challenges\nin their training and inference. Removing model components is perceived as a\nsolution to tackle the large model sizes, however, existing pruning methods\nsolely focus on performance, without considering an essential aspect for the\nresponsible use of LLMs: model fairness. It is crucial to address the fairness\nof LLMs towards diverse groups, such as women, Black people, LGBTQ+, Jewish\ncommunities, among others, as they are being deployed and available to a wide\naudience. In this work, first, we investigate how attention heads impact\nfairness and performance in pre-trained transformer-based language models. We\nthen propose a novel method to prune the attention heads that negatively impact\nfairness while retaining the heads critical for performance, i.e. language\nmodeling capabilities. Our approach is practical in terms of time and\nresources, as it does not require fine-tuning the final pruned, and fairer,\nmodel. Our findings demonstrate a reduction in gender bias by 19%, 19.5%,\n39.5%, 34.7%, 23%, and 8% for DistilGPT-2, GPT-2, GPT-Neo of two different\nsizes, GPT-J, and Llama 2 models, respectively, in comparison to the biased\nmodel, with only a slight decrease in performance.", + "authors": "Abdelrahman Zayed, Goncalo Mordido, Samira Shabanian, Ioana Baldini, Sarath Chandar", + "published": "2023-12-24", + "updated": "2023-12-24", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.CY", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.14473v1", + "title": "The Ethics of ChatGPT in Medicine and Healthcare: A Systematic Review on Large Language Models (LLMs)", + "abstract": "With the introduction of ChatGPT, Large Language Models (LLMs) have received\nenormous attention in healthcare. Despite their potential benefits, researchers\nhave underscored various ethical implications. While individual instances have\ndrawn much attention, the debate lacks a systematic overview of practical\napplications currently researched and ethical issues connected to them. Against\nthis background, this work aims to map the ethical landscape surrounding the\ncurrent stage of deployment of LLMs in medicine and healthcare. Electronic\ndatabases and preprint servers were queried using a comprehensive search\nstrategy. Studies were screened and extracted following a modified rapid review\napproach. Methodological quality was assessed using a hybrid approach. For 53\nrecords, a meta-aggregative synthesis was performed. Four fields of\napplications emerged and testify to a vivid exploration phase. Advantages of\nusing LLMs are attributed to their capacity in data analysis, personalized\ninformation provisioning, support in decision-making, mitigating information\nloss and enhancing information accessibility. However, we also identifies\nrecurrent ethical concerns connected to fairness, bias, non-maleficence,\ntransparency, and privacy. A distinctive concern is the tendency to produce\nharmful misinformation or convincingly but inaccurate content. A recurrent plea\nfor ethical guidance and human oversight is evident. Given the variety of use\ncases, it is suggested that the ethical guidance debate be reframed to focus on\ndefining what constitutes acceptable human oversight across the spectrum of\napplications. This involves considering diverse settings, varying potentials\nfor harm, and different acceptable thresholds for performance and certainty in\nhealthcare. In addition, a critical inquiry is necessary to determine the\nextent to which the current experimental use of LLMs is necessary and\njustified.", + "authors": "Joschka Haltaufderheide, Robert Ranisch", + "published": "2024-03-21", + "updated": "2024-03-21", + "primary_cat": "cs.CY", + "cats": [ + "cs.CY" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2305.07609v3", + "title": "Is ChatGPT Fair for Recommendation? Evaluating Fairness in Large Language Model Recommendation", + "abstract": "The remarkable achievements of Large Language Models (LLMs) have led to the\nemergence of a novel recommendation paradigm -- Recommendation via LLM\n(RecLLM). Nevertheless, it is important to note that LLMs may contain social\nprejudices, and therefore, the fairness of recommendations made by RecLLM\nrequires further investigation. To avoid the potential risks of RecLLM, it is\nimperative to evaluate the fairness of RecLLM with respect to various sensitive\nattributes on the user side. Due to the differences between the RecLLM paradigm\nand the traditional recommendation paradigm, it is problematic to directly use\nthe fairness benchmark of traditional recommendation. To address the dilemma,\nwe propose a novel benchmark called Fairness of Recommendation via LLM\n(FaiRLLM). This benchmark comprises carefully crafted metrics and a dataset\nthat accounts for eight sensitive attributes1 in two recommendation scenarios:\nmusic and movies. By utilizing our FaiRLLM benchmark, we conducted an\nevaluation of ChatGPT and discovered that it still exhibits unfairness to some\nsensitive attributes when generating recommendations. Our code and dataset can\nbe found at https://github.com/jizhi-zhang/FaiRLLM.", + "authors": "Jizhi Zhang, Keqin Bao, Yang Zhang, Wenjie Wang, Fuli Feng, Xiangnan He", + "published": "2023-05-12", + "updated": "2023-10-17", + "primary_cat": "cs.IR", + "cats": [ + "cs.IR", + "cs.CL", + "cs.CY" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2401.11033v4", + "title": "FAIR Enough: How Can We Develop and Assess a FAIR-Compliant Dataset for Large Language Models' Training?", + "abstract": "The rapid evolution of Large Language Models (LLMs) highlights the necessity\nfor ethical considerations and data integrity in AI development, particularly\nemphasizing the role of FAIR (Findable, Accessible, Interoperable, Reusable)\ndata principles. While these principles are crucial for ethical data\nstewardship, their specific application in the context of LLM training data\nremains an under-explored area. This research gap is the focus of our study,\nwhich begins with an examination of existing literature to underline the\nimportance of FAIR principles in managing data for LLM training. Building upon\nthis, we propose a novel framework designed to integrate FAIR principles into\nthe LLM development lifecycle. A contribution of our work is the development of\na comprehensive checklist intended to guide researchers and developers in\napplying FAIR data principles consistently across the model development\nprocess. The utility and effectiveness of our framework are validated through a\ncase study on creating a FAIR-compliant dataset aimed at detecting and\nmitigating biases in LLMs. We present this framework to the community as a tool\nto foster the creation of technologically advanced, ethically grounded, and\nsocially responsible AI models.", + "authors": "Shaina Raza, Shardul Ghuge, Chen Ding, Elham Dolatabadi, Deval Pandya", + "published": "2024-01-19", + "updated": "2024-04-03", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.09447v2", + "title": "How Trustworthy are Open-Source LLMs? An Assessment under Malicious Demonstrations Shows their Vulnerabilities", + "abstract": "The rapid progress in open-source Large Language Models (LLMs) is\nsignificantly driving AI development forward. However, there is still a limited\nunderstanding of their trustworthiness. Deploying these models at scale without\nsufficient trustworthiness can pose significant risks, highlighting the need to\nuncover these issues promptly. In this work, we conduct an adversarial\nassessment of open-source LLMs on trustworthiness, scrutinizing them across\neight different aspects including toxicity, stereotypes, ethics, hallucination,\nfairness, sycophancy, privacy, and robustness against adversarial\ndemonstrations. We propose advCoU, an extended Chain of Utterances-based (CoU)\nprompting strategy by incorporating carefully crafted malicious demonstrations\nfor trustworthiness attack. Our extensive experiments encompass recent and\nrepresentative series of open-source LLMs, including Vicuna, MPT, Falcon,\nMistral, and Llama 2. The empirical outcomes underscore the efficacy of our\nattack strategy across diverse aspects. More interestingly, our result analysis\nreveals that models with superior performance in general NLP tasks do not\nalways have greater trustworthiness; in fact, larger models can be more\nvulnerable to attacks. Additionally, models that have undergone instruction\ntuning, focusing on instruction following, tend to be more susceptible,\nalthough fine-tuning LLMs for safety alignment proves effective in mitigating\nadversarial trustworthiness attacks.", + "authors": "Lingbo Mo, Boshi Wang, Muhao Chen, Huan Sun", + "published": "2023-11-15", + "updated": "2024-04-02", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.17916v2", + "title": "LLM-Resistant Math Word Problem Generation via Adversarial Attacks", + "abstract": "Large language models (LLMs) have significantly transformed the educational\nlandscape. As current plagiarism detection tools struggle to keep pace with\nLLMs' rapid advancements, the educational community faces the challenge of\nassessing students' true problem-solving abilities in the presence of LLMs. In\nthis work, we explore a new paradigm for ensuring fair evaluation -- generating\nadversarial examples which preserve the structure and difficulty of the\noriginal questions aimed for assessment, but are unsolvable by LLMs. Focusing\non the domain of math word problems, we leverage abstract syntax trees to\nstructurally generate adversarial examples that cause LLMs to produce incorrect\nanswers by simply editing the numeric values in the problems. We conduct\nexperiments on various open- and closed-source LLMs, quantitatively and\nqualitatively demonstrating that our method significantly degrades their math\nproblem-solving ability. We identify shared vulnerabilities among LLMs and\npropose a cost-effective approach to attack high-cost models. Additionally, we\nconduct automatic analysis on math problems and investigate the cause of\nfailure, offering a nuanced view into model's limitation.", + "authors": "Roy Xie, Chengxuan Huang, Junlin Wang, Bhuwan Dhingra", + "published": "2024-02-27", + "updated": "2024-03-30", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.08656v1", + "title": "Linear Cross-document Event Coreference Resolution with X-AMR", + "abstract": "Event Coreference Resolution (ECR) as a pairwise mention classification task\nis expensive both for automated systems and manual annotations. The task's\nquadratic difficulty is exacerbated when using Large Language Models (LLMs),\nmaking prompt engineering for ECR prohibitively costly. In this work, we\npropose a graphical representation of events, X-AMR, anchored around individual\nmentions using a \\textbf{cross}-document version of \\textbf{A}bstract\n\\textbf{M}eaning \\textbf{R}epresentation. We then linearize the ECR with a\nnovel multi-hop coreference algorithm over the event graphs. The event graphs\nsimplify ECR, making it a) LLM cost-effective, b) compositional and\ninterpretable, and c) easily annotated. For a fair assessment, we first enrich\nan existing ECR benchmark dataset with these event graphs using an\nannotator-friendly tool we introduce. Then, we employ GPT-4, the newest LLM by\nOpenAI, for these annotations. Finally, using the ECR algorithm, we assess\nGPT-4 against humans and analyze its limitations. Through this research, we aim\nto advance the state-of-the-art for efficient ECR and shed light on the\npotential shortcomings of current LLMs at this task. Code and annotations:\n\\url{https://github.com/ahmeshaf/gpt_coref}", + "authors": "Shafiuddin Rehan Ahmed, George Arthur Baker, Evi Judge, Michael Regan, Kristin Wright-Bettner, Martha Palmer, James H. Martin", + "published": "2024-03-25", + "updated": "2024-03-25", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.08517v1", + "title": "Online Safety Analysis for LLMs: a Benchmark, an Assessment, and a Path Forward", + "abstract": "While Large Language Models (LLMs) have seen widespread applications across\nnumerous fields, their limited interpretability poses concerns regarding their\nsafe operations from multiple aspects, e.g., truthfulness, robustness, and\nfairness. Recent research has started developing quality assurance methods for\nLLMs, introducing techniques such as offline detector-based or uncertainty\nestimation methods. However, these approaches predominantly concentrate on\npost-generation analysis, leaving the online safety analysis for LLMs during\nthe generation phase an unexplored area. To bridge this gap, we conduct in this\nwork a comprehensive evaluation of the effectiveness of existing online safety\nanalysis methods on LLMs. We begin with a pilot study that validates the\nfeasibility of detecting unsafe outputs in the early generation process.\nFollowing this, we establish the first publicly available benchmark of online\nsafety analysis for LLMs, including a broad spectrum of methods, models, tasks,\ndatasets, and evaluation metrics. Utilizing this benchmark, we extensively\nanalyze the performance of state-of-the-art online safety analysis methods on\nboth open-source and closed-source LLMs. This analysis reveals the strengths\nand weaknesses of individual methods and offers valuable insights into\nselecting the most appropriate method based on specific application scenarios\nand task requirements. Furthermore, we also explore the potential of using\nhybridization methods, i.e., combining multiple methods to derive a collective\nsafety conclusion, to enhance the efficacy of online safety analysis for LLMs.\nOur findings indicate a promising direction for the development of innovative\nand trustworthy quality assurance methodologies for LLMs, facilitating their\nreliable deployments across diverse domains.", + "authors": "Xuan Xie, Jiayang Song, Zhehua Zhou, Yuheng Huang, Da Song, Lei Ma", + "published": "2024-04-12", + "updated": "2024-04-12", + "primary_cat": "cs.SE", + "cats": [ + "cs.SE", + "cs.AI", + "cs.CL", + "cs.CR", + "cs.LG" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2305.13862v2", + "title": "A Trip Towards Fairness: Bias and De-Biasing in Large Language Models", + "abstract": "Cheap-to-Build Very Large-Language Models (CtB-LLMs) with affordable training\nare emerging as the next big revolution in natural language processing and\nunderstanding. These CtB-LLMs are democratizing access to trainable Very\nLarge-Language Models (VLLMs) and, thus, may represent the building blocks of\nmany NLP systems solving downstream tasks. Hence, a little or a large bias in\nCtB-LLMs may cause huge harm. In this paper, we performed a large investigation\nof the bias of three families of CtB-LLMs, and we showed that debiasing\ntechniques are effective and usable. Indeed, according to current tests, the\nLLaMA and the OPT families have an important bias in gender, race, religion,\nand profession. In contrast to the analysis for other LLMs, we discovered that\nbias depends not on the number of parameters but on the perplexity. Finally,\nthe debiasing of OPT using LoRA reduces bias up to 4.12 points in the\nnormalized stereotype score.", + "authors": "Leonardo Ranaldi, Elena Sofia Ruzzetti, Davide Venditti, Dario Onorati, Fabio Massimo Zanzotto", + "published": "2023-05-23", + "updated": "2023-08-29", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2311.03033v1", + "title": "Beyond Words: A Mathematical Framework for Interpreting Large Language Models", + "abstract": "Large language models (LLMs) are powerful AI tools that can generate and\ncomprehend natural language text and other complex information. However, the\nfield lacks a mathematical framework to systematically describe, compare and\nimprove LLMs. We propose Hex a framework that clarifies key terms and concepts\nin LLM research, such as hallucinations, alignment, self-verification and\nchain-of-thought reasoning. The Hex framework offers a precise and consistent\nway to characterize LLMs, identify their strengths and weaknesses, and\nintegrate new findings. Using Hex, we differentiate chain-of-thought reasoning\nfrom chain-of-thought prompting and establish the conditions under which they\nare equivalent. This distinction clarifies the basic assumptions behind\nchain-of-thought prompting and its implications for methods that use it, such\nas self-verification and prompt programming.\n Our goal is to provide a formal framework for LLMs that can help both\nresearchers and practitioners explore new possibilities for generative AI. We\ndo not claim to have a definitive solution, but rather a tool for opening up\nnew research avenues. We argue that our formal definitions and results are\ncrucial for advancing the discussion on how to build generative AI systems that\nare safe, reliable, fair and robust, especially in domains like healthcare and\nsoftware engineering.", + "authors": "Javier Gonz\u00e1lez, Aditya V. Nori", + "published": "2023-11-06", + "updated": "2023-11-06", + "primary_cat": "cs.LG", + "cats": [ + "cs.LG", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.19465v1", + "title": "Towards Tracing Trustworthiness Dynamics: Revisiting Pre-training Period of Large Language Models", + "abstract": "Ensuring the trustworthiness of large language models (LLMs) is crucial. Most\nstudies concentrate on fully pre-trained LLMs to better understand and improve\nLLMs' trustworthiness. In this paper, to reveal the untapped potential of\npre-training, we pioneer the exploration of LLMs' trustworthiness during this\nperiod, focusing on five key dimensions: reliability, privacy, toxicity,\nfairness, and robustness. To begin with, we apply linear probing to LLMs. The\nhigh probing accuracy suggests that \\textit{LLMs in early pre-training can\nalready distinguish concepts in each trustworthiness dimension}. Therefore, to\nfurther uncover the hidden possibilities of pre-training, we extract steering\nvectors from a LLM's pre-training checkpoints to enhance the LLM's\ntrustworthiness. Finally, inspired by~\\citet{choi2023understanding} that mutual\ninformation estimation is bounded by linear probing accuracy, we also probe\nLLMs with mutual information to investigate the dynamics of trustworthiness\nduring pre-training. We are the first to observe a similar two-phase\nphenomenon: fitting and compression~\\citep{shwartz2017opening}. This research\nprovides an initial exploration of trustworthiness modeling during LLM\npre-training, seeking to unveil new insights and spur further developments in\nthe field. We will make our code publicly accessible at\n\\url{https://github.com/ChnQ/TracingLLM}.", + "authors": "Chen Qian, Jie Zhang, Wei Yao, Dongrui Liu, Zhenfei Yin, Yu Qiao, Yong Liu, Jing Shao", + "published": "2024-02-29", + "updated": "2024-02-29", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.13343v1", + "title": "Challenges and Contributing Factors in the Utilization of Large Language Models (LLMs)", + "abstract": "With the development of large language models (LLMs) like the GPT series,\ntheir widespread use across various application scenarios presents a myriad of\nchallenges. This review initially explores the issue of domain specificity,\nwhere LLMs may struggle to provide precise answers to specialized questions\nwithin niche fields. The problem of knowledge forgetting arises as these LLMs\nmight find it hard to balance old and new information. The knowledge repetition\nphenomenon reveals that sometimes LLMs might deliver overly mechanized\nresponses, lacking depth and originality. Furthermore, knowledge illusion\ndescribes situations where LLMs might provide answers that seem insightful but\nare actually superficial, while knowledge toxicity focuses on harmful or biased\ninformation outputs. These challenges underscore problems in the training data\nand algorithmic design of LLMs. To address these issues, it's suggested to\ndiversify training data, fine-tune models, enhance transparency and\ninterpretability, and incorporate ethics and fairness training. Future\ntechnological trends might lean towards iterative methodologies, multimodal\nlearning, model personalization and customization, and real-time learning and\nfeedback mechanisms. In conclusion, future LLMs should prioritize fairness,\ntransparency, and ethics, ensuring they uphold high moral and ethical standards\nwhen serving humanity.", + "authors": "Xiaoliang Chen, Liangbin Li, Le Chang, Yunhe Huang, Yuxuan Zhao, Yuxiao Zhang, Dinuo Li", + "published": "2023-10-20", + "updated": "2023-10-20", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2403.02839v1", + "title": "An Empirical Study of LLM-as-a-Judge for LLM Evaluation: Fine-tuned Judge Models are Task-specific Classifiers", + "abstract": "Recently, there has been a growing trend of utilizing Large Language Model\n(LLM) to evaluate the quality of other LLMs. Many studies have employed\nproprietary close-source models, especially GPT4, as the evaluator.\nAlternatively, other works have fine-tuned judge models based on open-source\nLLMs as the evaluator. In this study, we conduct an empirical study of\ndifferent judge models on their evaluation capability. Our findings indicate\nthat although the fine-tuned judge models achieve high accuracy on in-domain\ntest sets, even surpassing GPT4, they are inherently task-specific classifiers,\nand their generalizability and fairness severely underperform GPT4.", + "authors": "Hui Huang, Yingqi Qu, Jing Liu, Muyun Yang, Tiejun Zhao", + "published": "2024-03-05", + "updated": "2024-03-05", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2310.18333v3", + "title": "She had Cobalt Blue Eyes: Prompt Testing to Create Aligned and Sustainable Language Models", + "abstract": "As the use of large language models (LLMs) increases within society, as does\nthe risk of their misuse. Appropriate safeguards must be in place to ensure LLM\noutputs uphold the ethical standards of society, highlighting the positive role\nthat artificial intelligence technologies can have. Recent events indicate\nethical concerns around conventionally trained LLMs, leading to overall unsafe\nuser experiences. This motivates our research question: how do we ensure LLM\nalignment? In this work, we introduce a test suite of unique prompts to foster\nthe development of aligned LLMs that are fair, safe, and robust. We show that\nprompting LLMs at every step of the development pipeline, including data\ncuration, pre-training, and fine-tuning, will result in an overall more\nresponsible model. Our test suite evaluates outputs from four state-of-the-art\nlanguage models: GPT-3.5, GPT-4, OPT, and LLaMA-2. The assessment presented in\nthis paper highlights a gap between societal alignment and the capabilities of\ncurrent LLMs. Additionally, implementing a test suite such as ours lowers the\nenvironmental overhead of making models safe and fair.", + "authors": "Veronica Chatrath, Oluwanifemi Bamgbose, Shaina Raza", + "published": "2023-10-20", + "updated": "2023-12-15", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2402.07688v1", + "title": "CyberMetric: A Benchmark Dataset for Evaluating Large Language Models Knowledge in Cybersecurity", + "abstract": "Large Language Models (LLMs) excel across various domains, from computer\nvision to medical diagnostics. However, understanding the diverse landscape of\ncybersecurity, encompassing cryptography, reverse engineering, and managerial\nfacets like risk assessment, presents a challenge, even for human experts. In\nthis paper, we introduce CyberMetric, a benchmark dataset comprising 10,000\nquestions sourced from standards, certifications, research papers, books, and\nother publications in the cybersecurity domain. The questions are created\nthrough a collaborative process, i.e., merging expert knowledge with LLMs,\nincluding GPT-3.5 and Falcon-180B. Human experts spent over 200 hours verifying\ntheir accuracy and relevance. Beyond assessing LLMs' knowledge, the dataset's\nmain goal is to facilitate a fair comparison between humans and different LLMs\nin cybersecurity. To achieve this, we carefully selected 80 questions covering\na wide range of topics within cybersecurity and involved 30 participants of\ndiverse expertise levels, facilitating a comprehensive comparison between human\nand machine intelligence in this area. The findings revealed that LLMs\noutperformed humans in almost every aspect of cybersecurity.", + "authors": "Norbert Tihanyi, Mohamed Amine Ferrag, Ridhi Jain, Merouane Debbah", + "published": "2024-02-12", + "updated": "2024-02-12", + "primary_cat": "cs.AI", + "cats": [ + "cs.AI", + "cs.CR" + ], + "category": "LLM Fairness" + }, + { + "url": "http://arxiv.org/abs/2404.01349v1", + "title": "Fairness in Large Language Models: A Taxonomic Survey", + "abstract": "Large Language Models (LLMs) have demonstrated remarkable success across\nvarious domains. However, despite their promising performance in numerous\nreal-world applications, most of these algorithms lack fairness considerations.\nConsequently, they may lead to discriminatory outcomes against certain\ncommunities, particularly marginalized populations, prompting extensive study\nin fair LLMs. On the other hand, fairness in LLMs, in contrast to fairness in\ntraditional machine learning, entails exclusive backgrounds, taxonomies, and\nfulfillment techniques. To this end, this survey presents a comprehensive\noverview of recent advances in the existing literature concerning fair LLMs.\nSpecifically, a brief introduction to LLMs is provided, followed by an analysis\nof factors contributing to bias in LLMs. Additionally, the concept of fairness\nin LLMs is discussed categorically, summarizing metrics for evaluating bias in\nLLMs and existing algorithms for promoting fairness. Furthermore, resources for\nevaluating bias in LLMs, including toolkits and datasets, are summarized.\nFinally, existing research challenges and open questions are discussed.", + "authors": "Zhibo Chu, Zichong Wang, Wenbin Zhang", + "published": "2024-03-31", + "updated": "2024-03-31", + "primary_cat": "cs.CL", + "cats": [ + "cs.CL", + "cs.AI" + ], + "category": "LLM Fairness" + } +] \ No newline at end of file