{"id":"W7106823613","doi":"10.48448/140v-nv60","title":"Towards a Holistic and Automated Evaluation Framework for Multi-Level Comprehension of LLMs in Book-Length Contexts","year":2025,"lang":"","type":"other","venue":"Open MIND","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"","keywords":"Comprehension; Automatic summarization; Hierarchy; Narrative; Code (set theory); Recall; Sample (material)","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":[],"consensus_categories":[],"category_scores_codex":[0.01757946,0.003897112,0.00158411,0.009051132,0.001287079,0.008651712,0.003495797,0.003191315,0.01118385],"category_scores_gemma":[0.0804752,0.001091974,0.001817509,0.003655424,0.001480075,0.01254529,0.00783599,0.004323904,0.01030139],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002749893,"about_ca_system_score_gemma":0.004035093,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01470864,"about_ca_topic_score_gemma":0.0256378,"domain_scores_codex":[0.977975,0.01239498,0.001530092,0.00367439,0.003783946,0.0006416576],"domain_scores_gemma":[0.9532173,0.02628284,0.002107477,0.007217947,0.009679548,0.001494863],"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":[0.001662708,0.001246464,0.03531889,0.003401072,0.001177119,0.0007481403,0.007073061,0.05504316,0.03423218,0.02662301,0.2098337,0.6236404],"study_design_scores_gemma":[0.0002992728,0.0008445696,0.01613737,0.000670846,0.0002934937,0.0005812209,0.003782562,0.7836851,0.03106144,0.06686032,0.09541798,0.0003658549],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0754125,0.003404922,0.7147114,0.002736842,0.0004859517,0.001798273,0.04162203,0.141625,0.01820308],"genre_scores_gemma":[0.2629416,0.0006181111,0.6253826,0.00101273,0.0002478971,0.001586752,0.09701288,0.006907789,0.00428977],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01757946,"threshold_uncertainty_score":0.09297019,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2596008577206175,"score_gpt":0.4645924650690927,"score_spread":0.2049916073484752,"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."}}