{"id":"W4404782438","doi":"10.18653/v1/2024.emnlp-main.551","title":"StablePrompt : Automatic Prompt Tuning using Reinforcement Learning for Large Language Model","year":2024,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"Institute for Information and Communications Technology Promotion; Ministry of Science and ICT, South Korea","keywords":"Reinforcement learning; Computer science; Language model; Artificial intelligence","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":[0.0005984439,0.0001416953,0.0001493905,0.0001385017,0.0001857749,0.0004004691,0.0003916084,0.000047045,0.00003594369],"category_scores_gemma":[0.00004158204,0.0001254084,0.00007276061,0.0002095417,0.000005587143,0.0006584433,0.0002876272,0.0001580651,0.00001718307],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001075281,"about_ca_system_score_gemma":0.000150792,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003029007,"about_ca_topic_score_gemma":0.000003763508,"domain_scores_codex":[0.9985918,0.0000221431,0.0002754626,0.0004229922,0.0002445372,0.0004430118],"domain_scores_gemma":[0.9994253,0.00006626765,0.00004126655,0.0003656005,0.00003959925,0.00006199389],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[9.480269e-7,0.00001235228,0.0000199987,0.0002945541,0.00002746795,0.0000111889,0.006049516,0.832159,0.005491956,0.1367288,0.00009393338,0.01911027],"study_design_scores_gemma":[0.0001628717,0.00002595067,0.000001062047,0.0001360103,0.000008334094,0.000007166916,0.0001944169,0.996585,0.0008117409,0.001395291,0.0005029476,0.0001692241],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02432507,0.0002337404,0.971678,0.0001345732,0.0001914868,0.0003698125,5.836166e-7,0.0009729757,0.002093796],"genre_scores_gemma":[0.6591155,8.986083e-7,0.3378717,0.0001128235,0.00004758595,0.00002939373,0.000002526814,0.0000154082,0.002804174],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6347904,"threshold_uncertainty_score":0.5114008,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03792502520895687,"score_gpt":0.2995942561215851,"score_spread":0.2616692309126282,"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."}}