{"id":"W4389524581","doi":"10.18653/v1/2023.findings-emnlp.1036","title":"Can Retriever-Augmented Language Models Reason? The Blame Game Between the Retriever and the Language Model","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":8,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Computer science; Language model; Natural language processing; Artificial intelligence; Metric (unit)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.007044787,0.001816241,0.001377981,0.001313342,0.0006472308,0.00382274,0.004120414,0.00226033,0.008813233],"category_scores_gemma":[0.02862152,0.001164635,0.002839566,0.0008945339,0.001635615,0.01705889,0.003703034,0.004510927,0.005767064],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001672745,"about_ca_system_score_gemma":0.002221999,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01097072,"about_ca_topic_score_gemma":0.01962783,"domain_scores_codex":[0.9958248,0.00185739,0.0002911654,0.001017777,0.0007799931,0.000228957],"domain_scores_gemma":[0.9878407,0.006805899,0.0004676896,0.003558127,0.001080383,0.0002471395],"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":[0.001428615,0.000664949,0.006058557,0.001808483,0.0007627758,0.0007390272,0.002420389,0.1887931,0.02139702,0.07843211,0.0431679,0.6543271],"study_design_scores_gemma":[0.0001125826,0.0002098984,0.000572754,0.0001030945,0.0001771105,0.000308431,0.0004724252,0.8804141,0.01042539,0.08851306,0.01859503,0.0000960556],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05439804,0.001446689,0.9041462,0.004202715,0.0002635312,0.0003247483,0.002026421,0.02402195,0.009169634],"genre_scores_gemma":[0.4135685,0.0008584955,0.5673403,0.002392868,0.0001599482,0.0003324442,0.006104356,0.001332972,0.007910192],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01097072,"threshold_uncertainty_score":0.0372569,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02341281661982967,"score_gpt":0.2525640409814111,"score_spread":0.2291512243615814,"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."}}