{"id":"W4389520703","doi":"10.18653/v1/2023.emnlp-main.489","title":"TheoremQA: A Theorem-driven Question Answering Dataset","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Benchmark (surveying); Computer science; Domain (mathematical analysis); Code (set theory); Question answering; Quality (philosophy); Baseline (sea); Artificial intelligence; Programming language; Mathematics; Epistemology","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002627133,0.002267354,0.0006707227,0.002807246,0.001118136,0.001655443,0.003836731,0.004292777,0.01214796],"category_scores_gemma":[0.01341549,0.0004559928,0.002169047,0.002090588,0.0009325983,0.003019901,0.002266423,0.002905851,0.01077661],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002643663,"about_ca_system_score_gemma":0.002233428,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01962773,"about_ca_topic_score_gemma":0.03896876,"domain_scores_codex":[0.9973924,0.0007051654,0.0002839573,0.0007384653,0.0007238085,0.0001561198],"domain_scores_gemma":[0.9939016,0.003154127,0.0002994018,0.001022415,0.001173581,0.0004488313],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0006163075,0.0007871939,0.0119591,0.003149869,0.0002252541,0.0005539394,0.0006478885,0.01330296,0.005146747,0.007747923,0.904152,0.05171083],"study_design_scores_gemma":[0.001199492,0.0006924663,0.03020879,0.0004893105,0.0001518193,0.001330709,0.001120955,0.1489032,0.01350287,0.01871851,0.7834764,0.0002054616],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.05857262,0.002990791,0.01816713,0.003168987,0.0005993716,0.0009776126,0.8729107,0.02648913,0.01612372],"genre_scores_gemma":[0.04438863,0.0003074625,0.02535624,0.0008411408,0.00007660884,0.0006344652,0.9247033,0.0004427868,0.003249307],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.01962773,"threshold_uncertainty_score":0.04063898,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0263743278939783,"score_gpt":0.2841081159769848,"score_spread":0.2577337880830065,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}