{"id":"W6966678284","doi":"10.48448/vttz-0r79","title":"Collaborative Performance Prediction for Large Language Models","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Performance prediction; Downstream (manufacturing); Language model; Scaling; Predictive modelling; Scaling law","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","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001274737,0.0004005616,0.0003473349,0.001272174,0.0002301754,0.0002810834,0.0007224798,0.0002504196,0.0004753281],"category_scores_gemma":[0.000102501,0.0003488467,0.00006406812,0.002497377,0.0005509429,0.0006461393,0.0002016554,0.0003155409,0.00396539],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000432015,"about_ca_system_score_gemma":0.001082232,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000410159,"about_ca_topic_score_gemma":0.0003959864,"domain_scores_codex":[0.9969039,0.00002815662,0.0003096445,0.001029034,0.0009564387,0.0007728285],"domain_scores_gemma":[0.9986222,0.00003032788,0.000239418,0.0006175756,0.000321114,0.0001693137],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002711015,0.0001121314,0.00001274398,0.0003219527,0.00006045778,0.000005950548,0.001801701,0.000529814,0.001729371,0.01905117,0.9750625,0.001285103],"study_design_scores_gemma":[0.0005481779,0.000184832,0.000004219059,0.0004567417,0.00009297011,0.000006537174,0.001198003,0.5524733,0.00059277,0.001090787,0.4429139,0.0004377798],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.000419469,0.004022663,0.01911387,0.0001335529,0.002631702,0.002813914,0.01875021,0.002496365,0.9496183],"genre_scores_gemma":[0.05048661,0.0001274953,0.01465806,0.0001489609,0.001336345,0.0003474284,0.0006548429,0.001571445,0.9306688],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.5519435,"threshold_uncertainty_score":0.9998963,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01690932362608183,"score_gpt":0.3071650045323348,"score_spread":0.290255680906253,"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."}}