{"id":"W4309663019","doi":"10.1126/science.ade9097","title":"Human-level play in the game of<i>Diplomacy</i>by combining language models with strategic reasoning","year":2022,"lang":"en","type":"article","venue":"Science","topic":"Topic Modeling","field":"Computer Science","cited_by":201,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Cicero; Negotiation; Diplomacy; Computer science; Competition (biology); Reinforcement learning; Artificial intelligence; League; Natural language; Political science; Politics; Law; History; Ecology","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":[0.001721488,0.00009215444,0.0001154258,0.0001251456,0.0003700852,0.0001435663,0.002589412,0.00001168604,0.000007185532],"category_scores_gemma":[0.00001320662,0.00006778302,0.00001907886,0.00126747,0.0001942734,0.0006012787,0.0004850242,0.0002574844,9.885169e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005645066,"about_ca_system_score_gemma":0.000217464,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001794714,"about_ca_topic_score_gemma":0.00001934388,"domain_scores_codex":[0.9980585,0.0000959963,0.0001863541,0.0004240098,0.0008982774,0.0003368519],"domain_scores_gemma":[0.9991354,0.00005183514,0.0001059549,0.0006310791,0.00003377226,0.00004190709],"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":[0.000003998109,0.00009763975,0.0007061253,0.000009126736,0.000003138276,0.00005622999,0.03646035,0.1683172,0.02784741,0.7649925,0.00002866722,0.001477659],"study_design_scores_gemma":[0.0005226801,0.0001987006,0.001089365,0.00004149609,0.000002899643,0.00007160387,0.007176582,0.9620315,0.001619933,0.02697628,0.00002727919,0.0002416958],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8913401,0.0001540653,0.1004324,0.0001464688,0.00006188685,0.000140815,0.000003990444,0.00004456172,0.007675705],"genre_scores_gemma":[0.9878967,6.960827e-7,0.01176836,0.0002219404,0.000007553554,0.00002266356,7.191097e-7,0.000004455955,0.00007693394],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7937143,"threshold_uncertainty_score":0.4811815,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05109947060723923,"score_gpt":0.2899594067233681,"score_spread":0.2388599361161289,"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."}}