{"id":"W7106793789","doi":"10.48448/hdc5-t218","title":"When Format Changes Meaning: Investigating Semantic Inconsistency of Large Language Models","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":"Natural language; Language understanding; Phenomenon; Semantics (computer science); Semantic role labeling; Disk formatting","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":[0.006518422,0.0006699294,0.0007443904,0.0008103819,0.0006979234,0.002626147,0.002433578,0.001593669,0.001474627],"category_scores_gemma":[0.0503693,0.0005590381,0.0007668662,0.0007379108,0.002032399,0.00539784,0.002963878,0.002816438,0.0003469205],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001383461,"about_ca_system_score_gemma":0.001395899,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003634208,"about_ca_topic_score_gemma":0.003414904,"domain_scores_codex":[0.9961358,0.002089596,0.000182478,0.0006010911,0.0008173641,0.000173672],"domain_scores_gemma":[0.980261,0.01295105,0.001570737,0.003797784,0.001083788,0.0003355244],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.000936301,0.0003029676,0.03164858,0.0005449617,0.0004815438,0.001223296,0.00307686,0.6406981,0.01816914,0.1321229,0.0103444,0.1604509],"study_design_scores_gemma":[0.00002246792,0.00005684104,0.0008849804,0.00002810422,0.00003632674,0.00008821386,0.000200705,0.9100047,0.006057736,0.08084635,0.00174835,0.00002526606],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.3697217,0.0009189334,0.6132089,0.003807956,0.0001821078,0.0000967967,0.0009004947,0.00637564,0.004787344],"genre_scores_gemma":[0.933575,0.0001754655,0.06375854,0.0003938186,0.0000350168,0.00005943473,0.0008221309,0.0005961696,0.0005844739],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006518422,"threshold_uncertainty_score":0.03447312,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05782889447115869,"score_gpt":0.3193976272712312,"score_spread":0.2615687328000725,"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."}}