{"id":"W4283792669","doi":"10.1609/aaai.v36i10.21386","title":"Supervising Model Attention with Human Explanations for Robust Natural Language Inference","year":2022,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Topic Modeling","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Azrieli Foundation","keywords":"Computer science; Premise; Inference; Artificial intelligence; Natural (archaeology); Task (project management); Punctuation; Natural language; Natural language processing; Focus (optics); Language model; Machine learning; Linguistics","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":[],"consensus_categories":[],"category_scores_codex":[0.0004186943,0.0001729717,0.0001810097,0.0001488778,0.0007136452,0.0002165438,0.001737283,0.00003329854,0.00002339774],"category_scores_gemma":[0.0001177476,0.0001369812,0.00008776646,0.000450907,0.00009186572,0.0004845046,0.0004758161,0.0003214006,0.000002953505],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008443703,"about_ca_system_score_gemma":0.0001049207,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004481687,"about_ca_topic_score_gemma":0.00002674291,"domain_scores_codex":[0.9983082,0.00001551495,0.000377422,0.00047336,0.0005299613,0.0002955243],"domain_scores_gemma":[0.9988801,0.00006942979,0.0002617701,0.0002983975,0.0004417326,0.00004853329],"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":[0.00003097721,0.00009509209,0.0001250989,0.00003144172,0.00001187383,3.122815e-7,0.003051518,0.0305151,0.04198745,0.9045956,0.00003700728,0.0195186],"study_design_scores_gemma":[0.00004561824,0.0001440161,0.0000434021,0.00006572605,0.00001101398,0.000004169866,0.001868499,0.9352983,0.02657384,0.03574998,0.000009907526,0.0001854852],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3346348,0.00001852681,0.660996,0.002201852,0.0002204248,0.0005484703,0.00001262856,0.0001112028,0.001256078],"genre_scores_gemma":[0.9666736,0.000001809904,0.03268597,0.0001491873,0.00003515055,0.0001293784,0.00000322357,0.00001229016,0.0003094222],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9047832,"threshold_uncertainty_score":0.5585929,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1128407143945447,"score_gpt":0.3157759058673034,"score_spread":0.2029351914727586,"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."}}