{"paper":{"title":"Evaluating the Performance of Large Language Models on GAOKAO Benchmark","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Large language models achieve competitive scores on the Chinese GAOKAO exam but vary widely by subject.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chunyang Li, Liang He, Xiaotian Zhang, Xipeng Qiu, Yi Zong, Zhengyu Ying","submitted_at":"2023-05-21T14:39:28Z","abstract_excerpt":"Large Language Models(LLMs) have demonstrated remarkable performance across various natural language processing tasks; however, how to comprehensively and accurately assess their performance becomes an urgent issue to be addressed. This paper introduces GAOKAO-Bench, an intuitive benchmark that employs questions from the Chinese GAOKAO examination as test samples, including both subjective and objective questions. To align with human examination methods, we design a method based on zero-shot settings to evaluate the performance of LLMs. With human evaluation, we obtain the converted total scor"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Our findings reveal that LLMs have achieved competitive scores in Chinese GAOKAO examination, while they exhibit significant performance disparities across various subjects.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That zero-shot prompting on GAOKAO questions produces answers whose quality can be fairly compared to human exam performance via human grading.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"LLMs achieve competitive scores on GAOKAO exam questions but display large performance gaps across subjects.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Large language models achieve competitive scores on the Chinese GAOKAO exam but vary widely by subject.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"df4ba602b5a9d223f86614794c90e9047c55cef1fe19b4099ce0dae61944b78d"},"source":{"id":"2305.12474","kind":"arxiv","version":3},"verdict":{"id":"91ab21aa-fd8c-4cc4-9200-d8fac4b26310","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T12:23:54.910620Z","strongest_claim":"Our findings reveal that LLMs have achieved competitive scores in Chinese GAOKAO examination, while they exhibit significant performance disparities across various subjects.","one_line_summary":"LLMs achieve competitive scores on GAOKAO exam questions but display large performance gaps across subjects.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That zero-shot prompting on GAOKAO questions produces answers whose quality can be fairly compared to human exam performance via human grading.","pith_extraction_headline":"Large language models achieve competitive scores on the Chinese GAOKAO exam but vary widely by subject."},"references":{"count":19,"sample":[{"doi":"","year":null,"title":"Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt","work_id":"7165db1f-9442-4e9a-8a0d-43246db1d50f","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":2,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":2016,"title":"Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models","work_id":"bb63abb3-0d50-4362-b97c-b5e725b03b39","ref_index":3,"cited_arxiv_id":"2206.04615","is_internal_anchor":true},{"doi":"","year":null,"title":"加强管理：对凤堰古梯田保护区内的 游客进行管理，设置必要的警示标志， 禁止破坏梯田、采摘植物等行为。同 时，加强对古建筑民居群落、古寨堡、 古庙宇、古堰渠、古塘坝等文物遗存的 保护，防止游客在参观过程中对这些文 物遗存造成损害。","work_id":"d2bc52d5-0e2a-4d73-969b-b1c554bb984b","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"推广科普：在凤堰古梯田保护区内设 置科普展板，向游客介绍梯田的历史、 文化和生态环境，提高游客的文化素养 和环保意识，减少游客对梯田的破坏。","work_id":"9092f07e-a3ec-40f2-992c-08aa61280c2b","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":19,"snapshot_sha256":"b4dbbfe8698ddc22faeb70b9c09b1445219cfc9f134f6d507ff9023e4f4dc5e8","internal_anchors":2},"formal_canon":{"evidence_count":2,"snapshot_sha256":"eaf4a1530db43077596a47ebb42222e7b5856c83cba5f6c88f5b8ab7ea7e71ff"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}