{"id":"W6910667010","doi":"10.48448/wk1a-qv46","title":"Can Large Language Models Fix Data Annotation Errors? An Empirical Study Using Debatepedia for Query-Focused Text Summarization","year":2023,"lang":"en","type":"other","venue":"Open MIND","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"","keywords":"Automatic summarization; Relevance (law); Task (project management); Empirical research; Language model; Sampling (signal processing); Quality (philosophy); Annotation","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01142979,0.0009827319,0.0008337589,0.00259489,0.00109079,0.001969955,0.001784283,0.001309157,0.002965224],"category_scores_gemma":[0.07681378,0.0003186855,0.001003358,0.003222976,0.001080055,0.004482312,0.00140127,0.002601907,0.002735774],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001154707,"about_ca_system_score_gemma":0.001214758,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00607428,"about_ca_topic_score_gemma":0.007679038,"domain_scores_codex":[0.9892752,0.007440487,0.0006660126,0.00149846,0.0008829555,0.0002370107],"domain_scores_gemma":[0.9176657,0.06743999,0.003035386,0.00783163,0.003336617,0.0006906425],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.004304599,0.002514027,0.1369664,0.0057766,0.00113442,0.001200511,0.007471948,0.06046849,0.01430587,0.0114321,0.1964942,0.5579309],"study_design_scores_gemma":[0.001258525,0.002144063,0.08367559,0.0006949282,0.0008070712,0.001962349,0.007877453,0.6882219,0.03048062,0.02321586,0.1592934,0.0003682898],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8281295,0.006673249,0.09062653,0.004333858,0.000570001,0.0008882125,0.04818966,0.01194329,0.008645644],"genre_scores_gemma":[0.8116913,0.0007512849,0.0763021,0.0008842648,0.0002552674,0.0007909643,0.1056127,0.0007907984,0.00292133],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9885702,"threshold_uncertainty_score":0.06044722,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1978774936590849,"score_gpt":0.4370140280177273,"score_spread":0.2391365343586425,"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."}}