{
 "body": "\nPart of the [FewCLUE](fewclue.md) family.\n\n## What it measures\n\nTNEWS is short-text classification of Chinese news headlines from the Toutiao platform (published\nbefore May 2018) into one of 15 categories such as finance, sports, travel, education and\nentertainment. It is the CLUE benchmark's own TNEWS task, resampled into FewCLUE's few-shot format;\nthe authors additionally filtered the source headline pool by cross-validation to remove examples a\nsimple model could already classify easily, so the few-shot version stays discriminative between\nmodels rather than becoming trivially solvable.\n\n## Reading the numbers\n\nThe labelled public test set holds 2,010 headlines, scored by accuracy against a roughly 6.7%\nfifteen-way baseline (matching the near-uniform class split the paper reports). Human evaluators\nscored 71.0%, well below their scores on the two-way tasks -- fifteen categories are genuinely hard\nto discriminate even for people; the paper's best few-shot method (PET) reached only 54.5%. The\npaper's own stability ablation used exactly this task: resampling TNEWS's training/validation split\nunder an identical method swung accuracy by several points, so a single-split TNEWS score should be\nread cautiously rather than as a precise estimate. See the [FewCLUE](fewclue.md) family page for the\nshared few-shot protocol and contamination notes that apply here too.\n",
 "build": {
  "built_at": "2026-09-09T16:56:50+00:00",
  "commit": "0a599558854c0e238c03a0f0d725239cb28f9d11",
  "eligibility_as_of": "2026-09-09"
 },
 "disposition": {
  "canonical_id": "fewclue_tnews",
  "reasons": [],
  "status": "unassessed",
  "verified_results": []
 },
 "models_covered": [],
 "page": {
  "category": "reasoning",
  "dataset": {
   "languages": [
    "zh"
   ],
   "license": "",
   "modalities": [
    "text"
   ],
   "public_test_set": true,
   "size": 2010,
   "size_note": "2,010 labelled public test headlines (test_public.json, used for scoring), plus 1,500 in the private test set (original leaderboard only), 240 train and 240 dev examples per split (5 parallel splits, train_0..train_4 / dev_0..dev_4, plus a merged train_few_all/dev_few_all), and 20,000 unlabelled headlines not used for scoring. Figures from the FewCLUE paper's Table 1, matching the GitHub README. This is the CLUE benchmark's own TNEWS task, resampled into FewCLUE's few-shot splits; the source pool was filtered by cross-validation to remove headlines a simple model could already classify easily.\n",
   "splits": "train_0..train_4 (240 each) + train_few_all; dev_0..dev_4 (240 each) + dev_few_all; test_public (2,010, labelled); test (1,500, private); unlabeled (20,000)",
   "url": "https://github.com/CLUEbenchmark/FewCLUE/tree/main/datasets/tnews"
  },
  "freshness": {
   "researched": "2026-09-08",
   "researched_by": "sonnet-5 agent, batch 5, slice B"
  },
  "harness": {
   "opencompass": "FewCLUE_tnews"
  },
  "id": "fewclue_tnews",
  "leaderboard_url": "https://www.cluebenchmarks.com/fewclue.html",
  "lineage": {
   "family": "fewclue"
  },
  "measures": "A short Chinese news headline from the Toutiao platform; the model classifies it into one of 15 topic categories such as finance, sports or education, learned few-shot from 240 labelled training headlines (16 per category).\n",
  "metric": {
   "baseline_note": "6.7% is the paper's majority-class baseline, close to the fifteen-way uniform-random rate of about 6.7%, indicating a near-uniform class split in the test set. See the fewclue family page for the top overall few-shot method scores.\n",
   "direction": "higher_is_better",
   "human_baseline": 71.0,
   "max_score": 100,
   "name": "accuracy",
   "random_baseline": 6.7,
   "unit": "%"
  },
  "name": "FewCLUE: TNEWS (Short News Classification)",
  "page_kind": "subset",
  "paper": {
   "arxiv": "2107.07498",
   "title": "FewCLUE: A Chinese Few-shot Learning Evaluation Benchmark",
   "url": "https://arxiv.org/abs/2107.07498",
   "year": 2021
  },
  "publisher": {
   "authors": [
    "Liang Xu",
    "Xiaojing Lu",
    "Chenyang Yuan",
    "Xuanwei Zhang",
    "Huilin Xu",
    "Hu Yuan",
    "Guoao Wei",
    "Xiang Pan",
    "Xin Tian",
    "Libo Qin",
    "Hu Hai"
   ],
   "org": "CLUE team",
   "url": "https://github.com/CLUEbenchmark/FewCLUE"
  },
  "released": "2021-04",
  "repo_url": "https://github.com/CLUEbenchmark/FewCLUE",
  "sources": [
   {
    "accessed": "2026-09-08",
    "title": "FewCLUE: A Chinese Few-shot Learning Evaluation Benchmark (Xu et al., arXiv:2107.07498)",
    "url": "https://arxiv.org/abs/2107.07498"
   },
   {
    "accessed": "2026-09-08",
    "title": "CLUEbenchmark/FewCLUE GitHub repository (task description, dataset statistics)",
    "url": "https://github.com/CLUEbenchmark/FewCLUE"
   },
   {
    "accessed": "2026-09-08",
    "title": "OpenCompass FewCLUE_tnews dataset config",
    "url": "https://raw.githubusercontent.com/open-compass/opencompass/main/opencompass/configs/datasets/FewCLUE_tnews/FewCLUE_tnews_gen.py"
   }
  ],
  "status": "unknown",
  "subcategory": "short Chinese news headline topic classification (15-way), few-shot",
  "summary": "FewCLUE's short news classification task: sort a Chinese Toutiao headline into one of 15 categories, learned from 240 labelled training examples.",
  "tags": [
   "chinese",
   "few-shot",
   "fewclue-subset",
   "classification"
  ],
  "task_format": "Fifteen-way topic classification, graded on the single correct category; evaluated from a 240-example few-shot training split (16 per class), one of five parallel splits FewCLUE provides for this task.\n"
 }
}