{"id":"W3187467055","doi":"10.1145/3546577","title":"Post-hoc Interpretability for Neural NLP: A Survey","year":2022,"lang":"en","type":"review","venue":"ACM Computing Surveys","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":192,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Interpretability; Computer science; Categorization; Artificial intelligence; Post hoc; Machine learning; Data science; Accountability; Medicine","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":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","open_science"],"consensus_categories":[],"category_scores_codex":[0.02267497,0.0008798552,0.002336799,0.0004063899,0.0007641572,0.0006684168,0.009256629,0.0003047591,0.000101305],"category_scores_gemma":[0.01225553,0.0008527988,0.00109212,0.001806181,0.0001418606,0.0004584639,0.00626866,0.00109108,0.0001337641],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005009563,"about_ca_system_score_gemma":0.0007843031,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001738315,"about_ca_topic_score_gemma":0.0007400788,"domain_scores_codex":[0.9837036,0.01055293,0.001762411,0.002062411,0.0006589691,0.001259651],"domain_scores_gemma":[0.970855,0.02312321,0.0009869301,0.004257787,0.0005246151,0.0002524694],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000006275165,0.0001017709,0.0001019035,0.001995891,0.0000821646,0.00001594912,0.0003109762,0.0000859437,1.43758e-7,0.0005231227,0.0003198607,0.996456],"study_design_scores_gemma":[0.0001468017,0.0008566947,0.0004955076,0.001186734,0.0001159869,0.00008884116,0.00003865868,0.0457655,0.000004083316,0.001216739,0.9482879,0.001796568],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0001644114,0.7628838,0.230177,0.0000770268,0.003786387,0.001927196,0.0002638861,0.0006046657,0.0001156],"genre_scores_gemma":[0.00250143,0.9627492,0.03109632,0.0004985003,0.0006399711,0.000445609,0.00137392,0.000322486,0.000372504],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9946594,"threshold_uncertainty_score":0.9993923,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1804175306382501,"score_gpt":0.3942971125104791,"score_spread":0.213879581872229,"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."}}