{"id":"W4416466606","doi":"10.1038/s43588-025-00913-7","title":"Viability of using LLMs as models of human language processing","year":2025,"lang":"en","type":"article","venue":"Nature Computational Science","topic":"Neurobiology of Language and Bilingualism","field":"Neuroscience","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Human language; Language model; Language technology; Universal Networking Language; On Language; Computational linguistics; Work (physics)","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":[],"consensus_categories":[],"category_scores_codex":[0.0004129779,0.0000791971,0.0001423032,0.0001944816,0.0002098038,0.0000184754,0.0004511205,0.00007939043,0.000009930211],"category_scores_gemma":[0.0007864116,0.00006646609,0.0000410667,0.001068359,0.001012106,0.000233504,0.0001273754,0.0002128162,5.088988e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003010537,"about_ca_system_score_gemma":0.0003931704,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000331261,"about_ca_topic_score_gemma":0.000002477729,"domain_scores_codex":[0.9988076,0.00005632134,0.0002325247,0.0003684283,0.0003808081,0.0001543008],"domain_scores_gemma":[0.999192,0.0002618323,0.0001496502,0.0001450431,0.0002214876,0.00003001133],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.000008229247,0.00006229111,0.0002940641,0.00004227023,7.771218e-7,0.000009893079,0.0006341716,0.013166,0.9722014,0.01218145,0.000002468537,0.00139699],"study_design_scores_gemma":[0.0002023883,0.00004241335,0.001427439,0.00007731348,0.000007727959,0.00003618575,0.0001294849,0.07236552,0.8701373,0.05547202,0.00000393533,0.00009823638],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9953163,0.0001649288,0.00207064,0.0001001456,0.000116685,0.0001214364,0.000008448987,0.00002299437,0.002078387],"genre_scores_gemma":[0.9966521,3.581777e-7,0.002707345,0.0005971919,0.00001471962,8.048228e-7,0.000001393316,0.000002669497,0.00002347037],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1020641,"threshold_uncertainty_score":0.3729146,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02732870010255197,"score_gpt":0.3824924493057228,"score_spread":0.3551637492031708,"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."}}