{"id":"W4389524398","doi":"10.18653/v1/2023.emnlp-main.63","title":"Tree of Clarifications: Answering Ambiguous Questions with Retrieval-Augmented Large Language Models","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"Ministry of Science and ICT, South Korea; National Research Foundation of Korea; Institute for Information and Communications Technology Promotion; Electronics and Telecommunications Research Institute; National Research Foundation","keywords":"Computer science; Question answering; Ambiguity; Tree (set theory); Set (abstract data type); Artificial intelligence; Code (set theory); Natural language processing; Language model; Information retrieval; Domain (mathematical analysis); Machine learning; Mathematics; Programming language","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":[0.003325836,0.001964272,0.001106844,0.001760739,0.000795236,0.002052464,0.002444349,0.002724583,0.006436651],"category_scores_gemma":[0.01455101,0.0006723711,0.00161205,0.001160508,0.0008229296,0.005972543,0.002967298,0.003755874,0.004691799],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001285588,"about_ca_system_score_gemma":0.00145006,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008621345,"about_ca_topic_score_gemma":0.01743824,"domain_scores_codex":[0.9975745,0.001341408,0.00009138921,0.0006166805,0.0002672457,0.0001088211],"domain_scores_gemma":[0.9941248,0.004045393,0.0002276425,0.0008696466,0.0005102671,0.0002222491],"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":[0.001138221,0.0006370793,0.003643566,0.001293586,0.0003330586,0.0006309994,0.003392561,0.09828797,0.019628,0.02570833,0.1129793,0.7323273],"study_design_scores_gemma":[0.0001184949,0.0001864653,0.000751478,0.0001011488,0.00009298378,0.0002396905,0.0006190651,0.9129184,0.00567178,0.05261184,0.02661362,0.00007501336],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03482705,0.003792085,0.9118403,0.001870217,0.0004373449,0.0006543353,0.003799727,0.03699948,0.005779497],"genre_scores_gemma":[0.3069578,0.0008777498,0.6665732,0.001619989,0.0003144203,0.0006058039,0.01354819,0.001324973,0.008177892],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008621345,"threshold_uncertainty_score":0.02153271,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02744474896684697,"score_gpt":0.2701363584174316,"score_spread":0.2426916094505846,"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."}}