{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001706457,0.0002115003,0.0002332815,0.00006110792,0.000320759,0.0003170781,0.00178706,0.00008829595,0.00001117518],"category_scores_gemma":[0.0001547747,0.00009364738,0.00009535062,0.0008478755,0.0002059931,0.0002923963,0.0009903212,0.0003958208,0.00002671577],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000429246,"about_ca_system_score_gemma":0.00007856246,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001099564,"about_ca_topic_score_gemma":0.0001302477,"domain_scores_codex":[0.997927,0.0002376842,0.0002983705,0.0004821785,0.0005908305,0.0004639447],"domain_scores_gemma":[0.9974967,0.0006606505,0.0000986553,0.001601335,0.00005582791,0.00008680738],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001036445,0.00004352248,0.001003283,0.00004335625,0.0003718768,0.00009291094,0.1860649,0.07383245,0.0006019917,0.6467521,0.0187633,0.07232668],"study_design_scores_gemma":[0.0006136497,0.00001100858,0.0003511114,0.00001392424,0.00002395207,0.000007062309,0.001538896,0.9791291,0.0002724178,0.01773701,0.0001502946,0.0001516164],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.4193095,0.0006961866,0.5202821,0.05185587,0.0001682609,0.0008410083,0.00002755933,0.0006877459,0.006131826],"genre_scores_gemma":[0.9882928,0.00007620451,0.001519716,0.002033111,0.0001424,0.00002751213,0.000006512703,0.00002062122,0.007881125],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9052966,"threshold_uncertainty_score":0.3818829,"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."}}