{"id":"W6910185236","doi":"10.48448/cnfq-c089","title":"Detecting and Mitigating Hallucinations in Multilingual Summarisation","year":2023,"lang":"en","type":"other","venue":"Open MIND","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Hallucinating; Code (set theory); Reliability (semiconductor); Language model; Range (aeronautics); Artificial neural network; Transfer of learning","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":[0.004240816,0.002137403,0.001223027,0.002541887,0.0007223943,0.002190689,0.00164044,0.001433123,0.003083874],"category_scores_gemma":[0.02743283,0.0004293334,0.0009058659,0.001404961,0.0008016657,0.003868141,0.003699081,0.001703979,0.002455951],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007061482,"about_ca_system_score_gemma":0.000717258,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002710341,"about_ca_topic_score_gemma":0.005691123,"domain_scores_codex":[0.9969192,0.001284221,0.000354124,0.0006906188,0.0005673211,0.0001845569],"domain_scores_gemma":[0.9895037,0.005228786,0.0009662361,0.002070585,0.001879117,0.0003515875],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002395089,0.0003840661,0.009273157,0.002721661,0.0007441118,0.0008711955,0.001869409,0.05832146,0.06035393,0.002317491,0.04051355,0.8202348],"study_design_scores_gemma":[0.0004074732,0.002346734,0.01432048,0.0004137178,0.0006728455,0.001516192,0.001997826,0.8162735,0.1112747,0.01701807,0.03349147,0.0002670745],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4399768,0.01028263,0.4772749,0.001834649,0.001099505,0.0008277562,0.009703087,0.04799263,0.01100802],"genre_scores_gemma":[0.7475911,0.001278649,0.2156368,0.0005876604,0.0003737609,0.0003911262,0.02641222,0.001732962,0.005995587],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004240816,"threshold_uncertainty_score":0.0224278,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06237127909307904,"score_gpt":0.3641988333601937,"score_spread":0.3018275542671146,"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."}}