{"id":"W4402683454","doi":"10.18653/v1/2024.repl4nlp-1.21","title":"Whitening Not Recommended for Classification Tasks in LLMs","year":2024,"lang":"en","type":"article","venue":"","topic":"Artificial Intelligence in Law","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Windsor","funders":"","keywords":"Computer science","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":[],"consensus_categories":[],"category_scores_codex":[0.001008615,0.00005280964,0.00006927423,0.00008429482,0.000178949,0.0001558394,0.0001473443,0.00008403282,0.0005454811],"category_scores_gemma":[0.0003168766,0.00005017767,0.00004361743,0.0003271927,0.00008046342,0.000254465,0.0000131758,0.00008920459,0.0001787292],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001417937,"about_ca_system_score_gemma":0.00009832092,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001412532,"about_ca_topic_score_gemma":0.01064132,"domain_scores_codex":[0.9991661,0.00006243451,0.0002026169,0.0002012145,0.0001401395,0.0002274976],"domain_scores_gemma":[0.9993712,0.0004325224,0.00002037631,0.00008572015,0.00004622805,0.00004395864],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000008941826,0.00001607705,0.0003754228,0.000008190946,0.000003674671,0.000001013615,0.007739678,0.00001017562,0.001912806,0.8655131,0.005440968,0.11897],"study_design_scores_gemma":[0.00004858396,0.00004118665,0.0003872656,0.00007080966,0.000008795321,3.108456e-7,0.02552136,0.04016268,0.006842541,0.07066447,0.8560126,0.0002393272],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.0569103,0.0001591952,0.1253666,0.1123852,0.004245237,0.001617789,0.00001636413,0.0009263963,0.6983729],"genre_scores_gemma":[0.9896033,0.00002119181,0.002775404,0.0003181117,0.0002448875,0.00007518691,0.000004548695,0.000009174459,0.00694818],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.932693,"threshold_uncertainty_score":0.5972636,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2319611445318203,"score_gpt":0.4650844777137296,"score_spread":0.2331233331819093,"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."}}