license: apache-2.0
Japanese Stable LM Instruct Gamma 7B
Model Description
This is a 7B-parameter decoder-only Japanese language model fine-tuned on instruction-following datasets, built on top of the base model Japanese Stable LM Base Gamma 7B.
Performance
For our final model, we've used Stability AI Japan's Japanese MT-Bench as a more representative test of our model's capabilities. For our JA MT-Bench testing we use a Japanese prompt ("あなたは役立つアシスタントです。") as well as --num-choices 4 in an effort to reduce sampling variability, however we've still observed regular 0.5+ point (and sometimes even greater swings) between generations, as well as issues with default prompts and parameters when testing, so again, we'd urge caution in over-interpreting these scores and treating them as more of a probabilistic directional indicator, rather than a definitive score or ranking:
| Benchmark | Score |
|---|---|
| JA MT-Bench | 5.23 |
There is an JA-MT-Bench Leaderboard, for convenience, here is a comparison of the JA MT-Bench scores of some other models (our scores were rated by gpt-4-0613):
| Model | Score |
|---|---|
| gpt-4-0613 | 9.40 |
| gpt-4-1106-preview | 9.17 |
| gpt-3.5-turbo* | 8.41 |
| Qwen-72B-Chat | 7.97 |
| Qwen-14B-Chat | 7.47 |
| chatntq-ja-7b-v1.0 | 6.65 |
| Xwin-LM-70B-V0.1-GPTQ (q4-gs32-actorder) | 6.62 |
| shisa-gamma-7b-v1 | 6.12 |
| nekomata-14b-instruction (corrected prompt HF) | 5.57 |
| shisa-7B-v1-GPTQ (q4-gs32-actorder) | 5.35 |
| nekomata-14b-instruction (corrected prompt) | 5.30 |
| shisa-mega-7b-v1.2 | 5.27 |
| shisa-7b-v1 (full prompt) | 5.23 |
| Swallow-13b-instruct-hf | 5.17 |
| Swallow-70b-instruct-GPTQ (q4-gs32-actorder) | 5.15 |
| shisa-7b-v1 | 5.02 |
| shisa-7B-v1-AWQ (q4-gs128) | 4.78 |
| ELYZA-japanese-Llama-2-7b-fast-instruct* | 4.86 |
| shisa-bad-7b-v1 | 4.42 |
| Swallow-7b-instruct-hf | 4.21 |
| ja-stablelm-instruct-gamma-7b* | 4.01 |
| japanese-stablelm-instruct-alpha-7b* | 2.74 |
| Mistral-7B-OpenOrca-ja* | 2.23 |
| youri-7b-chat* | 2.00 |
| Mistral-7B-Instruct-v0.1* | 1.78 |
| llm-jp-13b-instruct-full-jaster-dolly-oasst-v1.0* | 1.31 |
| houou-instruction-7b-v1 | 1.02 |
| llm-jp-13b-instruct-full-jaster-dolly-oasst-v1.0 | 1.0 |
| llm-jp-13b-instruct-full-jaster-v1.0 | 1.0 |