{"id":"W7092316893","doi":"","title":"PRISM: Agentic Retrieval with LLMs for Multi-Hop Question Answering","year":2025,"lang":"","type":"article","venue":"ArXiv.org","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Question answering; Context (archaeology); Set (abstract data type); Document retrieval; Language model; Calibration","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":[0.003776167,0.001498384,0.001097824,0.002130095,0.0008534026,0.003025382,0.003018502,0.002274309,0.01133326],"category_scores_gemma":[0.01408016,0.0007562714,0.001683224,0.001378144,0.001074104,0.006867345,0.004558549,0.003071933,0.007523442],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001206178,"about_ca_system_score_gemma":0.002199674,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004012498,"about_ca_topic_score_gemma":0.006613654,"domain_scores_codex":[0.9973327,0.001237465,0.0002428663,0.000552727,0.0005384249,0.00009578839],"domain_scores_gemma":[0.9952206,0.00281718,0.0002310607,0.001041723,0.0005326368,0.0001566771],"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":[0.0008917846,0.0006798549,0.002509953,0.002544884,0.0004086081,0.0006077123,0.002090123,0.07226212,0.03654588,0.08866804,0.08656486,0.7062262],"study_design_scores_gemma":[0.0002238931,0.0002463622,0.0004212662,0.0001089398,0.0001586087,0.0003354891,0.0003780214,0.8023853,0.01890858,0.1046265,0.07211018,0.00009685611],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004075362,0.000908339,0.962701,0.0007062672,0.00009278463,0.0004261874,0.001035824,0.02681583,0.003238261],"genre_scores_gemma":[0.1069351,0.0005759411,0.8791821,0.000854573,0.0001462594,0.0007161802,0.00404287,0.001144827,0.006402077],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01133326,"threshold_uncertainty_score":0.0379135,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06427901365345795,"score_gpt":0.3066300447763392,"score_spread":0.2423510311228812,"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."}}