{"id":"W4286903770","doi":"10.48550/arxiv.2110.07752","title":"Hindsight: Posterior-guided training of retrievers for improved\\n open-ended generation","year":2021,"lang":"","type":"preprint","venue":"arXiv (Cornell University)","topic":"Topic Modeling","field":"Computer Science","cited_by":19,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Hindsight bias; Computer science; Context (archaeology); Generator (circuit theory); Relevance (law); Labrador Retriever; Artificial intelligence; Natural language processing; Information retrieval; Psychology; Cognitive psychology; Medicine; Power (physics); Geography","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001183736,0.0007090131,0.001222376,0.0004156466,0.000487663,0.0007970552,0.004647607,0.0007461197,0.00009719185],"category_scores_gemma":[0.0002561863,0.0009368192,0.0006398755,0.001104354,0.0001788441,0.001889486,0.005132813,0.0007173371,0.000005553071],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005705402,"about_ca_system_score_gemma":0.002103235,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004534562,"about_ca_topic_score_gemma":0.0001803791,"domain_scores_codex":[0.9944662,0.0003645399,0.001092192,0.003070095,0.0001908579,0.0008160789],"domain_scores_gemma":[0.9942773,0.0002094813,0.001354186,0.002603735,0.001199217,0.0003560816],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005726629,0.0006555967,0.0004019232,0.0008923917,0.001278987,0.0003649498,0.02006616,0.6901751,0.1102288,0.1386089,0.0004430658,0.03631144],"study_design_scores_gemma":[0.002612404,0.0002549508,0.00009552973,0.0002878376,0.0002346492,0.00002051613,0.001076392,0.9799185,0.01040598,0.003802395,0.000383177,0.000907608],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3595492,0.00005745815,0.6361073,0.0002152891,0.001890858,0.001266929,0.00004058173,0.00005540595,0.0008169809],"genre_scores_gemma":[0.9530227,0.0001426231,0.04376427,0.0002307444,0.0002628368,0.000004890836,0.0001037821,0.00004760198,0.002420579],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5934734,"threshold_uncertainty_score":0.9993082,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2562523096483355,"score_gpt":0.2391825804790763,"score_spread":0.01706972916925917,"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."}}