{"id":"W4408688975","doi":"10.32388/b69sky","title":"DAFE: LLM-Based Evaluation Through Dynamic Arbitration for Free-Form Question-Answering","year":2025,"lang":"en","type":"preprint","venue":"Qeios","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Arbitration; Question answering; Computer science; Information retrieval; Political science; 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"],"consensus_categories":[],"category_scores_codex":[0.0009925155,0.0002617914,0.0002645883,0.0001459441,0.0001533323,0.0002523761,0.001247906,0.0002929242,0.00001140585],"category_scores_gemma":[0.0003088728,0.0002874445,0.0001606954,0.0001583893,0.00001786522,0.0003679672,0.0005160818,0.000358042,0.000006628638],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004930512,"about_ca_system_score_gemma":0.000943316,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001970877,"about_ca_topic_score_gemma":0.0002367037,"domain_scores_codex":[0.9978026,0.00009748353,0.0004670916,0.0008200578,0.0005418716,0.0002708707],"domain_scores_gemma":[0.9976053,0.0001485332,0.0002478603,0.001578999,0.0003793544,0.00003997726],"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":[0.00002435076,0.0001934738,0.00007746035,0.001122527,0.00009494918,0.000002304531,0.001960222,0.6831188,0.0005572561,0.1925792,0.0006594774,0.11961],"study_design_scores_gemma":[0.0004341015,0.00003272668,0.00009172714,0.0003122515,0.00004452553,5.615422e-7,0.000007688326,0.8502818,0.0007365578,0.1472864,0.0005402482,0.0002313331],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003081862,0.0001981608,0.9877929,0.002378969,0.002272435,0.001633814,0.00003744348,0.0003314369,0.002272948],"genre_scores_gemma":[0.4439301,0.00001307398,0.5537965,0.0005614971,0.0002109931,0.0009756898,0.0002360146,0.00001789741,0.000258233],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4408483,"threshold_uncertainty_score":0.9999578,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03800586777782443,"score_gpt":0.3589599952477819,"score_spread":0.3209541274699575,"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."}}