{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001043705,0.0009776857,0.0006680239,0.0007805575,0.0004065961,0.001814156,0.001749123,0.0009088542,0.005904402],"category_scores_gemma":[0.007789243,0.0004431238,0.001216386,0.0005362897,0.0008896166,0.002993918,0.001662164,0.002036789,0.001717645],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001146978,"about_ca_system_score_gemma":0.001534369,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003325208,"about_ca_topic_score_gemma":0.007867237,"domain_scores_codex":[0.9994093,0.00016286,0.000047787,0.0001899137,0.0001495617,0.00004064311],"domain_scores_gemma":[0.997978,0.001187823,0.0001744676,0.0003856038,0.0001908713,0.00008325418],"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":[0.0003189299,0.0002727183,0.00242177,0.0004380949,0.0001829721,0.0002799669,0.0003269931,0.2958839,0.02505149,0.02944495,0.007999563,0.6373786],"study_design_scores_gemma":[0.00002221665,0.00003427832,0.0001735873,0.00002080771,0.00001775448,0.00004172234,0.00002569115,0.9565435,0.009144173,0.03217506,0.001789651,0.00001158149],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03187687,0.0002850934,0.9518341,0.0004931253,0.00008515779,0.0001253119,0.0004726546,0.01169844,0.003129235],"genre_scores_gemma":[0.3819647,0.0002607854,0.6117647,0.0004276774,0.00006002033,0.0002303973,0.001134768,0.000624799,0.003532029],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005904402,"threshold_uncertainty_score":0.0197522,"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."}}