{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/code/extract-answers","entry":"extract_answers","source":"Syntology graph, per-sample; not an archive number","read_at":"2026-09-24T18:15:14+00:00","claim":"Names are grouped by exact entry-name string. Same-named routines are NOT asserted to be equivalent; 'ran' means executed on a synthesized fixture, not correctness. n_samples_ran = sum of by_status over every status except 'unverified' (ran_draft_wrong and ran_fixture are failures of Syntology's instrument, not of the code); n_papers_ran = papers with at least one such sample.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"},"n_papers":10,"n_papers_ran":5,"units":"n_samples, n_samples_ran, n_samples_fingerprinted and by_status count distinct code bodies (code_sha256); n_places and n_places_pointer_only count places, one per (paper, code body) pair, which is also the unit of the samples list","n_samples":10,"n_samples_ran":5,"n_samples_fingerprinted":1,"n_places":10,"n_places_pointer_only":3,"by_status":{"ran_honours":0,"ran_violates":0,"ran_draft_wrong":2,"ran_fixture":0,"ran":3,"unverified":5},"syntology":{"atlas_url":null,"mcp":null,"mcp_per_sample":{"tool":"get_code","arguments_in":"samples[].mcp_get_code"},"developers":"https://syntology.ai/developers"},"samples":[{"arxiv_id":"2505.12929","paper":"/paper/do-not-let-low-probability-tokens-over","title":"Do Not Let Low-Probability Tokens Over-Dominate in RL for LLMs","date":"2025-05-19","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"zhyang2226/ar-lopti","path":"dataset_format.py","file_url":"https://github.com/zhyang2226/ar-lopti/blob/HEAD/dataset_format.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"13606d906fd1dfa3","mcp_get_code":{"code_sha256":"13606d906fd1dfa3"}},{"arxiv_id":"2409.17990","paper":"/paper/extracting-affect-aggregates-from","title":"Extracting Affect Aggregates from Longitudinal Social Media Data with Temporal Adapters for Large Language Models","date":"2024-09-26","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"dess-mannheim/temporal-adapters","path":"extract_answers.py","file_url":"https://github.com/dess-mannheim/temporal-adapters/blob/HEAD/extract_answers.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"50845729d38b2db5","mcp_get_code":{"code_sha256":"50845729d38b2db5"}},{"arxiv_id":"2407.06071","paper":"/paper/from-loops-to-oops-fallback-behaviors-of","title":"From Loops to Oops: Fallback Behaviors of Language Models Under Uncertainty","date":"2024-07-08","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"Mivg/fallbacks","path":"run_hallucinations_check.py","file_url":"https://github.com/Mivg/fallbacks/blob/HEAD/run_hallucinations_check.py","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"808e0a22c84796cd","mcp_get_code":{"code_sha256":"808e0a22c84796cd"}},{"arxiv_id":"2406.12288","paper":"/paper/an-investigation-of-neuron-activation-as-a","title":"An Investigation of Neuron Activation as a Unified Lens to Explain Chain-of-Thought Eliciting Arithmetic Reasoning of LLMs","date":"2024-06-18","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"dakingrai/neuron-analysis-cot-arithmetic-reasoning","path":"gsm8k_inference.py","file_url":"https://github.com/dakingrai/neuron-analysis-cot-arithmetic-reasoning/blob/HEAD/gsm8k_inference.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"86934d817edfb0b2","mcp_get_code":{"code_sha256":"86934d817edfb0b2"}},{"arxiv_id":"2404.10346","paper":"/paper/self-explore-to-avoid-the-pit-improving-the","title":"Self-Explore: Enhancing Mathematical Reasoning in Language Models with Fine-grained Rewards","date":"2024-04-16","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"hbin0701/Self-Explore","path":"gen/utils_others.py","file_url":"https://github.com/hbin0701/Self-Explore/blob/HEAD/gen/utils_others.py","status":"ran","verification_level":1,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"fc107f0a7cb2bb77","mcp_get_code":{"code_sha256":"fc107f0a7cb2bb77"}},{"arxiv_id":"2112.02268","paper":"/paper/bridging-pre-trained-models-and-downstream","title":"Bridging Pre-trained Models and Downstream Tasks for Source Code Understanding","date":"2021-12-04","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"wangdeze18/DACL","path":"algorithm_classification/evaluator/extract_answers.py","file_url":"https://github.com/wangdeze18/DACL/blob/HEAD/algorithm_classification/evaluator/extract_answers.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"bedbea7ed3a8588a","mcp_get_code":{"code_sha256":"bedbea7ed3a8588a"}},{"arxiv_id":"2109.12072","paper":"/paper/sd-qa-spoken-dialectal-question-answering-for","title":"SD-QA: Spoken Dialectal Question Answering for the Real World","date":"2021-09-24","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"ffaisal93/sd-qa","path":"extract_qa_from_jsonl.py","file_url":"https://github.com/ffaisal93/sd-qa/blob/HEAD/extract_qa_from_jsonl.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"code_sha256_prefix":"fd24ba8d7375d425","mcp_get_code":{"code_sha256":"fd24ba8d7375d425"}},{"arxiv_id":"1505.00468","paper":"/paper/vqa-visual-question-answering","title":"VQA: Visual Question Answering","date":"2015-05-03","month_inferred_from_arxiv_id":null,"title_source":"archive","repo":"chirag26495/DAN_VQA","path":"utils/build_vqa_inputs.py","file_url":"https://github.com/chirag26495/DAN_VQA/blob/HEAD/utils/build_vqa_inputs.py","status":"ran_draft_wrong","verification_level":1,"contract_check":"OUTPUT_MISDECLARED","metamorphic_tier":"deterministic","behaviour_fingerprint":false,"licence":"NONE","inline_ok":false,"code_sha256_prefix":"0cf6e2086607dd93","mcp_get_code":{"code_sha256":"0cf6e2086607dd93"}},{"arxiv_id":"2025.findings-acl.301","paper":null,"title":"arXiv:2025.findings-acl.301","date":null,"month_inferred_from_arxiv_id":null,"title_source":null,"repo":"OpenBMB/ConsJudge","path":"src/RAG_train/infer_acc.py","file_url":"https://github.com/OpenBMB/ConsJudge/blob/HEAD/src/RAG_train/infer_acc.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"4fa8544d540bdcd7","mcp_get_code":{"code_sha256":"4fa8544d540bdcd7"}},{"arxiv_id":"2025.emnlp-main.796","paper":null,"title":"arXiv:2025.emnlp-main.796","date":null,"month_inferred_from_arxiv_id":null,"title_source":null,"repo":"yukyunglee/CheckEval","path":"src/inference_checkeval.py","file_url":"https://github.com/yukyunglee/CheckEval/blob/HEAD/src/inference_checkeval.py","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"code_sha256_prefix":"dc1f5c8bd2851eb7","mcp_get_code":{"code_sha256":"dc1f5c8bd2851eb7"}}]}