{"id":"W4404399990","doi":"10.48550/arxiv.2411.08010","title":"ExpressivityBench: Can LLMs Communicate Implicitly?","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Biomedical Text Mining and Ontologies","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Atomic Energy of Canada Limited","keywords":"Computer science; Business; Computer security","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008736572,0.001219512,0.0005232093,0.001244089,0.00061619,0.004007959,0.001521862,0.001893324,0.004054887],"category_scores_gemma":[0.07036897,0.0004332863,0.0006622489,0.0008377084,0.001707825,0.008347612,0.00334424,0.00227511,0.002004293],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001014282,"about_ca_system_score_gemma":0.0009266086,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001590019,"about_ca_topic_score_gemma":0.00170419,"domain_scores_codex":[0.9879332,0.007543847,0.0007613644,0.001292717,0.002165299,0.0003034962],"domain_scores_gemma":[0.9500436,0.0366178,0.002701236,0.007135641,0.002954216,0.0005475847],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002768298,0.0008460491,0.04278572,0.005570173,0.0006636176,0.000979721,0.02022767,0.08919064,0.09460365,0.1110892,0.04569478,0.5855806],"study_design_scores_gemma":[0.0002229025,0.001289594,0.01249755,0.0007336136,0.0002930996,0.0008585847,0.005487267,0.6130294,0.08946393,0.177788,0.09800982,0.0003261531],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2502242,0.001210505,0.6951735,0.003633974,0.0003264222,0.0006003194,0.007145298,0.01828538,0.02340044],"genre_scores_gemma":[0.8439091,0.0003484873,0.1426571,0.0006926375,0.00007992088,0.0005061745,0.006615594,0.00125626,0.003934664],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008736572,"threshold_uncertainty_score":0.04620391,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08480621804238049,"score_gpt":0.2175119563183353,"score_spread":0.1327057382759548,"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."}}