{"id":"W4416035455","doi":"10.18653/v1/2025.emnlp-main.1556","title":"Improving Rule-based Reasoning in LLMs using Neurosymbolic Representations","year":2025,"lang":"","type":"article","venue":"","topic":"Logic, Reasoning, and Knowledge","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Air Force Office of Scientific Research; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Representation (politics); Process (computing); Key (lock); Frame (networking); Normative; Case-based reasoning","routes":{"ca_aff":true,"ca_fund":true,"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","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0009165009,0.0004957131,0.0005788039,0.0009628636,0.0007277403,0.001133581,0.001456233,0.0002550827,0.0001215654],"category_scores_gemma":[0.001143386,0.0005131509,0.0002686466,0.003699673,0.000199004,0.0009484621,0.0009366719,0.000649478,0.00006010839],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003375364,"about_ca_system_score_gemma":0.001652138,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003406614,"about_ca_topic_score_gemma":0.0004862588,"domain_scores_codex":[0.9954373,0.0004789426,0.0009834019,0.001566341,0.0004031266,0.001130876],"domain_scores_gemma":[0.9971949,0.0005203686,0.0003166105,0.001457965,0.0002789453,0.0002312151],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001162041,0.001695885,0.238468,0.0006333523,0.0001134102,0.000443722,0.005634164,0.03821526,0.03102991,0.43628,0.0008872706,0.2464827],"study_design_scores_gemma":[0.001398526,0.000055187,0.02284325,0.0003977124,0.00004730243,0.00001290493,0.0002020247,0.9664735,0.006207037,0.001622219,0.0002518801,0.0004884953],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.166423,0.001028718,0.7659339,0.0007401062,0.002645349,0.0006250688,0.000002878612,0.000272631,0.06232835],"genre_scores_gemma":[0.9572096,0.00002406274,0.03763017,0.0009329624,0.000153244,0.00002901971,0.000003351311,0.00002770232,0.003989848],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9282582,"threshold_uncertainty_score":0.9999033,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02318787494492101,"score_gpt":0.2958163936077655,"score_spread":0.2726285186628445,"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."}}