prompt
listlengths 1
1
| completion
listlengths 1
1
| answer
stringlengths 1
22.9k
| score
float64 0
1
| temperature
float64 0.7
0.7
| model
stringclasses 1
value | max_tokens
int64 4.1k
4.1k
|
|---|---|---|---|---|---|---|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n12.Β Β S3: /fa: rs/?",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"The line appears to be a malformed or garbled URL or code fragment, likely from a navigation or UI element.\", \n \"coherency_score\": 1,\n \"coherency_reason\": \"Extremely fragmented and lacks meaningful structure or grammatical coherence.\"\n}"
}
] |
12.Β Β S3: /fa: rs/?
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n 3. AB(SentPA(A) ββ§βSentPA(B) β (T[Aβ β§βB] β (TAβ β§βTB)))",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"The line contains complex logical notation unrelated to narrative or article content, typical of technical or mathematical expressions in UI elements or boilerplate.\",\n \"coherency_score\": 1,\n \"coherency_reason\": \"The text is fragmented and non-linguistic, consisting of symbols and operators with no clear grammatical or semantic structure.\"\n}"
}
] |
3. AB(SentPA(A) ββ§βSentPA(B) β (T[Aβ β§βB] β (TAβ β§βTB)))
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nSouth German, Swiss, and Austrian topographic name from dialect Au βwater meadowβ, βstreamβ (see AUE).",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is part of a linguistic explanation and appears to be main content from a descriptive or etymological text.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and clearly expresses a meaningful linguistic definition with proper grammar and structure.\"\n}"
}
] |
South German, Swiss, and Austrian topographic name from dialect Au βwater meadowβ, βstreamβ (see AUE).
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n figure; external appearance.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line refers to a descriptive element of an object, likely part of a product or article description.\", \n \"coherency_score\": 4,\n \"coherency_reason\": \"The phrase is clear and grammatically complete, though slightly awkward in structure. It reads like a natural snippet from a descriptive text.\"\n}"
}
] |
figure; external appearance.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nΒ© Copyright NewsWeekly.com.au 2017",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a copyright notice, which falls under footer boilerplate.\", \n \"coherency_score\": 5,\n \"coherency_reason\": \"The line is a complete, grammatically correct, and meaningful statement of copyright ownership.\"\n}"
}
] |
Β© Copyright NewsWeekly.com.au 2017
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nSnack - None",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This appears to be a placeholder or UI label, likely from a form or settings page indicating no selection.\", \n \"coherency_score\": 3,\n \"coherency_reason\": \"Partially coherent as it forms a recognizable phrase, but lacks context and is incomplete in meaning.\"\n}"
}
] |
Snack - None
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nacchi Β Β Β Β Β <unavailable>",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"Contains a placeholder or system message indicating unavailability, typical of UI or technical status indicators.\",\n \"coherency_score\": 2,\n \"coherency_reason\": \"Partially coherent as it contains recognizable text, but 'acchi' is nonsensical and '<unavailable>' is a fragmented status message with no clear meaning.\"\n}"
}
] |
acchi Β Β Β Β Β <unavailable>
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nby: James Tilbury",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a generic author attribution typically found in metadata or sidebars, not part of main article content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase 'by: James Tilbury' is a clear, well-formed attribution and reads naturally as a standard metadata line.\"\n}"
}
] |
by: James Tilbury
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nRauch also gives a list which is like the DSM of psychology: if you see two or three of these six symptoms, then someoneβs getting canceled:",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line is part of a narrative discussing psychological symptoms in a journalistic or analytical context.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as a complete and fluent statement.\"\n}"
}
] |
Rauch also gives a list which is like the DSM of psychology: if you see two or three of these six symptoms, then someoneβs getting canceled:
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nOther information on the label is given in percentages. Food contains fat, protein, carbohydrates, and fiber. Food also contains vitamins, such as A and C, and minerals, such as calcium and iron. Nutrition specialists know how much of each one kids and adults should get every day to have a healthy diet. The percent daily value on a food label tells you how this food can help someone meet these daily goals.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This text provides substantive, narrative information about food nutrition and daily values, forming part of a main article or product description.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentences are well-structured, fluent, and form a coherent explanation of nutrition labels and daily values.\"\n}"
}
] |
Other information on the label is given in percentages. Food contains fat, protein, carbohydrates, and fiber. Food also contains vitamins, such as A and C, and minerals, such as calcium and iron. Nutrition specialists know how much of each one kids and adults should get every day to have a healthy diet. The percent daily value on a food label tells you how this food can help someone meet these daily goals.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nI Love the Comic-Book-Like Storylines",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is a substantive statement about content style, likely part of a product or article description.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, grammatically correct, and reads like a natural, fluent expression of opinion.\"\n}"
}
] |
I Love the Comic-Book-Like Storylines
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n 38. 38",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"Numerical notation without context suggests a page number or pagination control, which is UI clutter.\",\n \"coherency_score\": 2,\n \"coherency_reason\": \"The line is incomplete and lacks meaningful context or sentence structure, making it partially coherent only as a standalone number.\"\n}"
}
] |
38. 38
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nSaturday, December 08, 2007",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a date stamp, commonly found in footers or metadata sections of web pages.\", \n \"coherency_score\": 5,\n \"coherency_reason\": \"The line is a clear, well-formed date in a standard format and reads naturally as a complete phrase.\"\n}"
}
] |
Saturday, December 08, 2007
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n2) Tech notes, Bear Power supplies",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This appears to be a navigation or sidebar label related to technical content, likely part of a menu or widget.\", \n \"coherency_score\": 4,\n \"coherency_reason\": \"The phrase is clear and grammatically complete, though slightly awkward in context. It reads like a label or heading rather than a full sentence.\"\n}"
}
] |
2) Tech notes, Bear Power supplies
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n 16. Congratulations Chrissy! Iβm so happy for you!",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is a personal congratulatory message that appears to be part of a narrative or article content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, grammatically correct, and reads like a natural, fluent piece of conversational text.\"\n}"
}
] |
16. Congratulations Chrissy! Iβm so happy for you!
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nBy the spring of 1636, John Cotton had become the focus of the other clergymen in the colony. Thomas Shepard, the minister of Newtown (later named Cambridge), wrote a letter to Cotton and warned him of the strange opinions being circulated among his Boston parishioners. Shepard also expressed concern about Cotton's preaching, and some of the points of his theology.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is narrative text describing historical events and religious concerns, clearly part of main content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-structured, and reads like a natural, fluent passage from a historical account.\"\n}"
}
] |
By the spring of 1636, John Cotton had become the focus of the other clergymen in the colony. Thomas Shepard, the minister of Newtown (later named Cambridge), wrote a letter to Cotton and warned him of the strange opinions being circulated among his Boston parishioners. Shepard also expressed concern about Cotton's preaching, and some of the points of his theology.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nMeet our Writer",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a promotional or introductory label typically found in navigation or account sections, not part of main article content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase is grammatically correct, clear, and reads like a natural, well-formed sentence.\"\n}"
}
] |
Meet our Writer
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nDaily Questions - November 19, 2020 by AutoModerator in femalefashionadvice",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This appears to be a post metadata line from a blog or article, indicating the date and author, which is part of the main content structure.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The line is fully coherent and grammatically correct, forming a natural part of article metadata.\"\n}"
}
] |
Daily Questions - November 19, 2020 by AutoModerator in femalefashionadvice
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nThis is an interesting story, but the fact is no less interesting, because of the commercial cleverness that saw and took instant advantage of the power of a more or less sentimental association. The real origin of \"mission\" furniture is this: A number of years ago a manufacturer made two very clumsy chairs, the legs of which were merely three-inch posts, the backs straight, and the whole construction crude to a degree. They were shown at a spring exhibition of furniture, where they attracted a good deal of attention as a novelty. It was just at the time that the California Missions were exciting much attention, and a clever Chicago dealer, seeing the advertising value that lay in the idea, bought both pieces and advertised them as having been found in the California Missions.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The text is a substantive narrative excerpt from an article about furniture history and marketing.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence structure is clear, fluent, and logically connected, forming a coherent and natural passage of explanatory text.\"\n}"
}
] |
This is an interesting story, but the fact is no less interesting, because of the commercial cleverness that saw and took instant advantage of the power of a more or less sentimental association. The real origin of "mission" furniture is this: A number of years ago a manufacturer made two very clumsy chairs, the legs of which were merely three-inch posts, the backs straight, and the whole construction crude to a degree. They were shown at a spring exhibition of furniture, where they attracted a good deal of attention as a novelty. It was just at the time that the California Missions were exciting much attention, and a clever Chicago dealer, seeing the advertising value that lay in the idea, bought both pieces and advertised them as having been found in the California Missions.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nGet new listings like these in your email",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a call to action promoting email subscriptions, typical of promotional banners or ads.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is clear, grammatically correct, and reads like a natural, well-formed phrase.\"\n}"
}
] |
Get new listings like these in your email
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nWho want with I? You want it with I?",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line appears to be part of a narrative or dialogue, possibly from a main content piece like a blog or article.\",\n \"coherency_score\": 2,\n \"coherency_reason\": \"The sentence is fragmented and grammatically incorrect, with unclear subject-verb relationships and repetition, making it partially coherent but awkward and incomplete.\"\n}"
}
] |
Who want with I? You want it with I?
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nSouth: Provides the most light; plants can be situated far into the room and still receive bright light.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This describes a factual observation about lighting and plant placement, which is part of substantive content related to interior design or gardening.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as a complete explanatory statement.\"\n}"
}
] |
South: Provides the most light; plants can be situated far into the room and still receive bright light.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nHow the Strawberry Laser Lipo Works",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is a substantive explanatory phrase that appears to be part of a product or article description.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase is fully coherent, well-formed, and reads naturally as a topic sentence in a product or technical explanation.\"\n}"
}
] |
How the Strawberry Laser Lipo Works
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n\"I know what you mean. We have to find a way to get into that pary. I hear the security at these parties is tough though. It will take every ounce of resourcefullness we have to break the ranks of security and blend in with the crowd\"",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The text is part of a narrative discussion about attending parties and overcoming security, which fits as main content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentences are well-structured, grammatically correct, and form a coherent, natural conversation flow.\"\n}"
}
] |
"I know what you mean. We have to find a way to get into that pary. I hear the security at these parties is tough though. It will take every ounce of resourcefullness we have to break the ranks of security and blend in with the crowd"
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nΪ©ΩΩ
Ψ§Ψͺ Ϊ©ΩΫΨ―Ϋ",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"Ϊ©ΩΩ
Ψ§Ψͺ Ϊ©ΩΫΨ―Ϋ is a common metadata tag used in SEO, typically found in sidebar widgets or content metadata sections.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase is grammatically correct and clearly forms a meaningful, well-formed phrase in Persian.\"\n}"
}
] |
Ϊ©ΩΩ
Ψ§Ψͺ Ϊ©ΩΫΨ―Ϋ
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nAnd, those who exhaust all Tiers should be able to get food stamps AND cash assistance even if they don't have any kids, when applying for public assistance.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is a substantive statement about public assistance eligibility, part of a narrative explanation.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as a complete thought.\"\n}"
}
] |
And, those who exhaust all Tiers should be able to get food stamps AND cash assistance even if they don't have any kids, when applying for public assistance.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nThis week, a KGOP radio host was appalled. βWhat could we possibly learn from Norway?β",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line is part of a narrative sentence that appears to be from a main article or blog post.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, grammatically correct, and reads naturally as a complete thought.\"\n}"
}
] |
This week, a KGOP radio host was appalled. βWhat could we possibly learn from Norway?β
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nGeneral board for soon-to-be 1Ls / Re: socratic method",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This appears to be part of a discussion or conversation in a forum or community, relevant to legal education content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, grammatically correct, and forms a natural, meaningful phrase in context.\"\n}"
}
] |
General board for soon-to-be 1Ls / Re: socratic method
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nALSO READ:Β Renewable energy to contribute 30% of electricity supply in Nigeria by 2030, says Minister",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line provides substantive information about renewable energy in Nigeria and is part of a narrative article.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as a complete, meaningful statement.\"\n}"
}
] |
ALSO READ:Β Renewable energy to contribute 30% of electricity supply in Nigeria by 2030, says Minister
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n β’ Alphabetize",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a UI control or filter option, typical of a sidebar or sorting interface.\", \n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase 'Alphabetize' is a clear, well-formed instruction and reads naturally as a standalone command.\"\n}"
}
] |
β’ Alphabetize
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n 1. I've been using STAR for all my RNA-Seq data for a few months now. I ran comparisons between STAR and Tophat 1.3.3-2.0.3, and STAR consistently yielded better results in a fraction of the time.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is substantive narrative text describing experience with RNA-Seq tools and performance comparisons.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as part of a scientific or technical discussion.\"\n}"
}
] |
1. I've been using STAR for all my RNA-Seq data for a few months now. I ran comparisons between STAR and Tophat 1.3.3-2.0.3, and STAR consistently yielded better results in a fraction of the time.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nIptyxβs Blessed Soil",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This appears to be a product or brand name related to soil, likely part of a main content description.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase is fully coherent and reads like a natural, well-formed product name or title.\"\n}"
}
] |
Iptyxβs Blessed Soil
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n sys_p = tfest(z2,2);",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a line of code, not natural language content, and appears to be part of a programming or technical script.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"The line is syntactically correct and follows a clear pattern in programming syntax, making it understandable as a code snippet despite not being natural language.\"\n}"
}
] |
sys_p = tfest(z2,2);
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nPlease enter your comment!",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a form prompt typically found in user interaction interfaces, not part of main article content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase is clear, grammatically correct, and reads like a natural instruction in a form context.\"\n}"
}
] |
Please enter your comment!
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nThe next question I put to her was pretty obvious, really - βIf the moon is female and the sun is male, what do you think would happen if they ever came to togetherβ?",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is a narrative question that appears to be part of a main article or blog post, not a UI element or promotional message.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as a complete and fluent question in a conversational context.\"\n}"
}
] |
The next question I put to her was pretty obvious, really - βIf the moon is female and the sun is male, what do you think would happen if they ever came to togetherβ?
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nA big thanks for your time.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is a sincere acknowledgment that fits within the context of main content, such as a blog post or article conclusion.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The line is fully coherent, grammatically correct, and reads like a natural, fluent sentence in a human-written text.\"\n}"
}
] |
A big thanks for your time.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n 1. Thanks for that, Joao. The term βsci-artβ can indeed be replaced by any of the others you mention (my use of that one is arbitrary), but the larger issue is why the genre feels compelled to separate itself from the rest of art. Letβs say, for example, that you make art that involves architectural imagery. Do you then call it βarch-artβ? Or you make portraits with heavy psychological overtones; does this make your work βpsy-artβ? Youβd probably just call it art, since the term is elastic enough to cover just about any kind of content. That it feels compelled to distinguish itself from other art implies, to me anyway, that the genre considers itself some kind of new hybrid form (i.e., part art, part science). This is a huge claim, because if thatβs the case we can reasonably expect it to involve some aspect of real science (as opposed to, say, poetic allusions to it, or an appropriation of its imagery for the workβs own purposes). So to me itβs not so much the language that matters but the claims, motives, and intentions implicit in it.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line is part of a substantive discussion on the nature of 'sci-art' and artistic classification, forming a coherent narrative within a main article or blog post.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The text is fully coherent, well-formed, and reads like a natural, fluent passage of explanatory prose with clear logical flow and meaningful content.\"\n}"
}
] |
1. Thanks for that, Joao. The term βsci-artβ can indeed be replaced by any of the others you mention (my use of that one is arbitrary), but the larger issue is why the genre feels compelled to separate itself from the rest of art. Letβs say, for example, that you make art that involves architectural imagery. Do you then call it βarch-artβ? Or you make portraits with heavy psychological overtones; does this make your work βpsy-artβ? Youβd probably just call it art, since the term is elastic enough to cover just about any kind of content. That it feels compelled to distinguish itself from other art implies, to me anyway, that the genre considers itself some kind of new hybrid form (i.e., part art, part science). This is a huge claim, because if thatβs the case we can reasonably expect it to involve some aspect of real science (as opposed to, say, poetic allusions to it, or an appropriation of its imagery for the workβs own purposes). So to me itβs not so much the language that matters but the claims, motives, and intentions implicit in it.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nDainon. said...",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This appears to be a user-generated comment or profile identifier, typical of social media or comment sections.\", \n \"coherency_score\": 2,\n \"coherency_reason\": \"Partially coherent but incomplete and lacks context or a clear subject; reads like a fragment.\"\n}"
}
] |
Dainon. said...
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nWhy is it called Relational and what does that word mean?",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line is part of a substantive explanatory narrative about terminology.\", \n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and naturally phrased with clear meaning.\"\n}"
}
] |
Why is it called Relational and what does that word mean?
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n β’ 3 Control Lines (BUSY, POUT, SELECT) (Pin 11, 12, 13)",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a technical pin configuration label, typical of circuit diagrams or hardware documentation, not part of main content.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"The line is clear and grammatically structured, describing hardware pins with consistent terminology, though slightly technical and disconnected from narrative flow.\"\n}"
}
] |
β’ 3 Control Lines (BUSY, POUT, SELECT) (Pin 11, 12, 13)
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nHe also stated that (while such broadcasts required wide transmissions) he was developing a means by which privacy of ground telephonic messages could be maintained among callers. This method of individuation would take place through the ground, insuring that no one could eavesdrop on anyoneβs conversations.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line is part of a substantive narrative discussing technical developments in telephonic privacy.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-structured, and reads like a natural, fluent excerpt from a technical or journalistic article.\"\n}"
}
] |
He also stated that (while such broadcasts required wide transmissions) he was developing a means by which privacy of ground telephonic messages could be maintained among callers. This method of individuation would take place through the ground, insuring that no one could eavesdrop on anyoneβs conversations.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nRed Lyin' in Winter?",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line appears to be part of a narrative or descriptive passage, possibly from a literary or poetic context.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase is fully coherent and well-formed, with a natural, fluent structure that reads like a complete and meaningful sentence.\"\n}"
}
] |
Red Lyin' in Winter?
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nwhole links, while the dogs bounced back and forth, rambling along the",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line contains narrative text describing a scene involving dogs and movement, which appears to be part of a main article or story.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"The sentence is mostly coherent and reads like a natural, flowing phrase, though it is slightly incomplete and lacks a clear ending.\"\n}"
}
] |
whole links, while the dogs bounced back and forth, rambling along the
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nLeave a Reply",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a generic UI action button label disconnected from article content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The phrase 'Leave a Reply' is a complete, well-formed sentence in natural language.\"\n}"
}
] |
Leave a Reply
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nStill, as successful as the book was, Lee and Goodman knew something was missing. There wasnβt anyone who had the ability of keeping the team together; there was no strong leader. They needed someone who had a long-established history to be part of the team. Enter Captain America; co-created in the 1940βs by writer Joe Simon and artist Jack Kirby. Capβs greatest run at that point, was in the 40βs, when he was popular as a patriotic superhero. But after WWII, sales dwindled for his comic books and he only appeared in comic books occasionally during the 1950s. His return to the Marvel Universe in issue 4 of The Avengers is not just a pivotal moment in Avengers history but in comic book history as well. This was because he is a pop icon who symbolized what was best about America and his heroics dated back to World War II.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line presents a narrative about Captain America's history and role in the Marvel Universe, which is part of main content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The text is fully coherent, well-structured, and reads like a natural, fluent passage from a blog or article.\"\n}"
}
] |
Still, as successful as the book was, Lee and Goodman knew something was missing. There wasnβt anyone who had the ability of keeping the team together; there was no strong leader. They needed someone who had a long-established history to be part of the team. Enter Captain America; co-created in the 1940βs by writer Joe Simon and artist Jack Kirby. Capβs greatest run at that point, was in the 40βs, when he was popular as a patriotic superhero. But after WWII, sales dwindled for his comic books and he only appeared in comic books occasionally during the 1950s. His return to the Marvel Universe in issue 4 of The Avengers is not just a pivotal moment in Avengers history but in comic book history as well. This was because he is a pop icon who symbolized what was best about America and his heroics dated back to World War II.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n 2. Anonymous11:24",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This appears to be a user timestamp or comment identifier, typical of a forum or discussion board footer or sidebar widget.\",\n \"coherency_score\": 3,\n \"coherency_reason\": \"Partially coherent as it follows a standard format but lacks complete context or meaning on its own.\"\n}"
}
] |
2. Anonymous11:24
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n Another problem is mixing drivers from different manufacturers. It's very unlikely that those, or even most/any, 12s and 10s will be equally loud/efficient.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line discusses technical aspects of audio drivers, which is part of substantive content likely from a product review or technical article.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is clear, grammatically correct, and presents a complete, meaningful idea with logical flow.\"\n}"
}
] |
Another problem is mixing drivers from different manufacturers. It's very unlikely that those, or even most/any, 12s and 10s will be equally loud/efficient.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n Mar 29 '17 at 13:55",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a timestamp, commonly found in comment sections or activity logs, which is part of UI metadata rather than main content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The timestamp is a clear, well-formed sequence of date and time in a standard format, readable and natural in context.\"\n}"
}
] |
Mar 29 '17 at 13:55
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nlog('Informational message here', 'info');",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a JavaScript log statement, typical of developer console output or debugging code, not part of main content.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"The line is a clear, grammatically correct syntax for a logging function, though it's technical and not natural language. It reads like a complete and meaningful code snippet.\"\n}"
}
] |
log('Informational message here', 'info');
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n β’ Water is where bacteria flourishes. Removing it from your beauty products, particularly ones which aren't using pump to dispense, can beΒ more hygienic.Β ",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line provides a substantive explanation about hygiene and beauty products, forming part of a coherent narrative.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as a complete and fluent statement.\"\n}"
}
] |
β’ Water is where bacteria flourishes. Removing it from your beauty products, particularly ones which aren't using pump to dispense, can beΒ more hygienic.Β
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nAnonymous said...",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a user comment placeholder typically found in comment sections or social media features, which falls under web clutter.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"The phrase is clear and grammatically complete, resembling a natural sentence structure used in comment sections.\"\n}"
}
] |
Anonymous said...
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nBrittney Griner also talks with a low voice and doesn't wear makeup while she plays, which means she looks a little different than her beponytailed teammates. Is this everyone's cue to go act like a dickhead about it on the internet?! YES! It is!",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line is part of a narrative discussion about a player's appearance and online reactions, which falls within the scope of main content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, grammatically correct, and reads naturally as a complete thought with clear meaning and flow.\"\n}"
}
] |
Brittney Griner also talks with a low voice and doesn't wear makeup while she plays, which means she looks a little different than her beponytailed teammates. Is this everyone's cue to go act like a dickhead about it on the internet?! YES! It is!
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nIt goes like this: Years ago a famous psychologist and his team took a group of five monkeys and placed them in a cage with a ladder and a bunch of bananas hanging down from the top of the cage in the middle of the space. The monkeys at first were allowed to eat some bananas and they loved climbing to the top and grabbing a banana.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is narrative text describing an experiment, clearly part of main content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-structured, and reads like a natural, fluent excerpt from a story or article.\"\n}"
}
] |
It goes like this: Years ago a famous psychologist and his team took a group of five monkeys and placed them in a cage with a ladder and a bunch of bananas hanging down from the top of the cage in the middle of the space. The monkeys at first were allowed to eat some bananas and they loved climbing to the top and grabbing a banana.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\n β’ Social services",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a sidebar widget or navigation item related to social services, typical of UI elements in web pages.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"The phrase is clear and grammatically complete, reading like a standard label in a menu or sidebar.\"\n}"
}
] |
β’ Social services
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nThe same could be achieved for cubic (i.e. no x^3),",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is a substantive sentence discussing mathematical concepts and appears to be part of an article or explanation.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, grammatically correct, and reads like a natural, fluent part of a technical explanation.\"\n}"
}
] |
The same could be achieved for cubic (i.e. no x^3),
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nI feel more like an individual producer, but I still enjoy making music with other people, even just for fun. I think I made music either with Dario and Marco, Skee Mask and Andrea, but recently I didnβt have too much time to invest in collaborations as I was very busy with the LP.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"This is narrative text describing personal experiences with music production and collaborations.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence structure is clear, fluent, and naturally flows as part of a coherent personal reflection.\"\n}"
}
] |
I feel more like an individual producer, but I still enjoy making music with other people, even just for fun. I think I made music either with Dario and Marco, Skee Mask and Andrea, but recently I didnβt have too much time to invest in collaborations as I was very busy with the LP.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\ncomments powered by Disqus",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a widget placeholder for a comment system, typical of web clutter.\", \n \"coherency_score\": 4,\n \"coherency_reason\": \"The phrase is clear and grammatically complete, though it's a technical placeholder rather than narrative content.\"\n}"
}
] |
comments powered by Disqus
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nGolf health and fitness tools is developed to assist a golf player develop toughness in specific muscle mass so regarding enhance their golf video game as well as reduced their ratings.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line discusses golf player development and fitness, which is part of substantive content related to a product or service.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"The sentence is mostly coherent and conveys a clear idea, though it contains awkward phrasing and minor grammatical issues.\"\n}"
}
] |
Golf health and fitness tools is developed to assist a golf player develop toughness in specific muscle mass so regarding enhance their golf video game as well as reduced their ratings.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nEbola Airport Checks Expand: Nurses Get Training.Β The federal government is closing a gap in Ebola screening at airports while states from New York to Texas to California work to get hospitals and nurses ready in case another patient turns up somewhere in the U.S. with the deadly disease.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line presents a substantive narrative about Ebola screening and healthcare preparedness, forming part of a main article's content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-formed, and reads naturally as a complete and fluent piece of journalistic text.\"\n}"
}
] |
Ebola Airport Checks Expand: Nurses Get Training.Β The federal government is closing a gap in Ebola screening at airports while states from New York to Texas to California work to get hospitals and nurses ready in case another patient turns up somewhere in the U.S. with the deadly disease.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nPreparation Time:Β ",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"This is a metadata label typically found in recipe or product information sections, serving as a UI control or header rather than substantive content.\",\n \"coherency_score\": 3,\n \"coherency_reason\": \"Partially coherent as a phrase, but it is incomplete and lacks a value, making it a fragment rather than a full sentence.\"\n}"
}
] |
Preparation Time:Β
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nSTR 24 (+7)",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": true,\n \"clutter_reason\": \"STR 24 (+7) appears to be a statistic or data point, likely from a sports or performance context, commonly found in UI elements like player stats or game summaries, which are often part of web clutter.\",\n \"coherency_score\": 3,\n \"coherency_reason\": \"The line is partially coherent as it follows a pattern typical in data displays, but lacks context or a complete sentence structure, making it slightly awkward and incomplete.\"\n}"
}
] |
STR 24 (+7)
| 0
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nKhurt said...",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The phrase appears to be part of a narrative or quote within main content, possibly from a blog or article.\",\n \"coherency_score\": 4,\n \"coherency_reason\": \"It is mostly coherent as a fragment of a sentence, suggesting a person is speaking or quoting. It is slightly incomplete but reads naturally in context.\"\n}"
}
] |
Khurt said...
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
[
{
"content": "\nYou are a classifier that analyzes a single line of text taken from a web page.\n\nYour task has two parts:\n\n1. Classify whether the line is web clutter or main content.\n - Web clutter includes:\n - Navigation links or menus (e.g., \"Home | About | Contact\")\n - Cookie banners and consent messages\n - Sign-up / login prompts and account bars\n - Ads, promotional banners, calls to action (e.g., \"Buy now\", \"Limited time offer\")\n - Footer boilerplate (e.g., copyright, privacy policy, terms of service)\n - Sidebar widgets and tag clouds\n - Pagination / UI controls (e.g., \"Next page\", \"1 2 3 4\", \"Share on Twitter\")\n - Search bars and filter/sort controls\n - Generic UI labels or button text disconnected from article content\n - Main content includes:\n - Sentences that appear to be part of the main article, blog post, product description,\n documentation body text, etc.\n - Substantive explanatory or narrative text, even if short.\n\n2. Rate the coherency of the line on a 0β5 scale.\n - Coherency measures how much the line reads like a complete, meaningful snippet of natural\n language, not how βgoodβ or βusefulβ it is.\n - Use this scale:\n - 0 β Nonsense, random characters, or garbled text.\n - 1 β Extremely fragmented; almost no meaning.\n - 2 β Partially coherent; clearly incomplete or heavily cut off.\n - 3 β Mostly coherent; understandable but may be slightly awkward or incomplete.\n - 4 β Clear and coherent; reads like a normal sentence or phrase.\n - 5 β Fully coherent and well-formed; fluent and natural.\nOutput\n\nRespond only with a single JSON object, no extra text, using this schema:\n\n{\n \"is_clutter\": true or false,\n \"clutter_reason\": \"<short explanation of why it is or isn't clutter>\",\n \"coherency_score\": 0-5,\n \"coherency_reason\": \"<short explanation of the coherency score>\"\n}\n\n- is_clutter must be true if the line is web clutter, otherwise false.\n- clutter_reason should be 1β2 short sentences.\n- coherency_score must be an integer from 0 to 5.\n- coherency_reason should briefly justify the score in 1β2 short sentences.\n- Do not include any commentary outside the JSON.\n\nLine:\n\nIll-fated attempts by 1st worlders to \"help\" third worlders create a viscous cycle, in which \"aid\" often creates dependency, corruption, and overcrowding which thereby creates (in theory) more guilt among first worlders who feel like they need to do even more to help out.",
"role": "user"
}
] |
[
{
"role": "assistant",
"content": "{\n \"is_clutter\": false,\n \"clutter_reason\": \"The line presents a substantive narrative about aid and its consequences, clearly part of an article or blog post's main content.\",\n \"coherency_score\": 5,\n \"coherency_reason\": \"The sentence is fully coherent, well-structured, and reads like a natural, fluent excerpt from a thoughtful analysis.\"\n}"
}
] |
Ill-fated attempts by 1st worlders to "help" third worlders create a viscous cycle, in which "aid" often creates dependency, corruption, and overcrowding which thereby creates (in theory) more guilt among first worlders who feel like they need to do even more to help out.
| 1
| 0.7
|
Qwen/Qwen3-4B-Instruct-2507
| 4,096
|
End of preview. Expand
in Data Studio
No dataset card yet
- Downloads last month
- 1