{"id":"W4387561465","doi":"10.48550/arxiv.2310.06786","title":"OpenWebMath: An Open Dataset of High-Quality Mathematical Web Text","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Mathematics, Computing, and Information Processing","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Fujitsu; Natural Sciences and Engineering Research Council of Canada; Government of Canada; Canadian Institute for Advanced Research; Brigham Young University","keywords":"Computer science; Language model; Quality (philosophy); Preprocessor; Information retrieval; Code (set theory); Natural language processing; Artificial intelligence; World Wide Web; Programming language","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.001841582,0.0003342469,0.0007159237,0.00028617,0.0002073639,0.0006420867,0.00713847,0.0002404373,0.0001330847],"category_scores_gemma":[0.0002007914,0.0003501398,0.0001027128,0.0006150167,0.0001453474,0.002152006,0.01094987,0.0004310447,0.0004068788],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000901718,"about_ca_system_score_gemma":0.0004379418,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002150379,"about_ca_topic_score_gemma":0.00002822339,"domain_scores_codex":[0.9974224,0.0002173712,0.0008691793,0.0009042912,0.0002295432,0.0003572138],"domain_scores_gemma":[0.9955871,0.0002629875,0.001120235,0.002543641,0.0002654925,0.0002206076],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001234948,0.0002566538,0.00006997532,0.001027504,0.00006994615,0.0000329363,0.001024758,0.01737257,0.00002334588,0.975633,0.003517602,0.0009593677],"study_design_scores_gemma":[0.0004063793,0.00004158944,0.0001615456,0.0002586154,0.00003172814,0.000004226534,0.0001836601,0.6449403,0.0001579688,0.3528575,0.0005626619,0.0003938154],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1184664,0.000005557701,0.8758798,0.00009586062,0.0003467329,0.0005059319,0.0004338514,0.0002974974,0.003968327],"genre_scores_gemma":[0.95448,0.0000253211,0.04399708,0.0001454776,0.00004391891,0.000001900692,0.0004626447,0.0000218231,0.0008218515],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8360136,"threshold_uncertainty_score":0.999895,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2422326316852966,"score_gpt":0.2815237572692198,"score_spread":0.03929112558392323,"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."}}