{"id":"W6891672098","doi":"10.48448/5f7b-mw39","title":"Precise Zero-Shot Dense Retrieval without Relevance Labels","year":2022,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Relevance (law); Embedding; Encoding (memory); Relevance feedback; Document retrieval; Encoder; Vector space model; Bottleneck","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.003254671,0.0008849808,0.0008933494,0.001994704,0.0007787061,0.0003491551,0.003537737,0.0003622819,0.02060232],"category_scores_gemma":[0.002476469,0.0008659543,0.0001504299,0.005459006,0.003345358,0.0004934071,0.001386087,0.001377075,0.009488476],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00158533,"about_ca_system_score_gemma":0.002578782,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002282547,"about_ca_topic_score_gemma":0.0005138587,"domain_scores_codex":[0.9907739,0.0002690956,0.0007368824,0.002417912,0.004174259,0.00162795],"domain_scores_gemma":[0.9952115,0.000240565,0.0009959689,0.00261631,0.000341383,0.0005942464],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002019399,0.000351992,0.0007921603,0.00008394733,0.00007076114,0.0001416312,0.0002675821,0.0002194314,0.02317955,0.002223291,0.9698936,0.002574079],"study_design_scores_gemma":[0.001050225,0.0002610527,0.00005926578,0.0002198279,0.000109792,0.0001079072,0.0001036416,0.002063141,0.001538758,0.001202529,0.9920024,0.001281446],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.003248194,0.003653938,0.0007512089,0.0003369206,0.003582425,0.00261473,0.001695661,0.003283151,0.9808338],"genre_scores_gemma":[0.01484009,0.0002420964,0.01831607,0.0006152781,0.0007304277,0.00006713472,0.0002556051,0.002599532,0.9623337],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.02210877,"threshold_uncertainty_score":0.9993791,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03680912418638017,"score_gpt":0.3184238519734348,"score_spread":0.2816147277870546,"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."}}