{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/adversarial-policies-beat-professional-level","title":"Adversarial Policies Beat Superhuman Go AIs","arxiv_id":"2211.00241","date":"2022-11-01","proceeding":null,"authors":["Tony T. Wang","Adam Gleave","Tom Tseng","Kellin Pelrine","Nora Belrose","Joseph Miller","Michael D. Dennis","Yawen Duan","Viktor Pogrebniak","Sergey Levine","Stuart Russell"],"abstract":"We attack the state-of-the-art Go-playing AI system KataGo by training adversarial policies against it, achieving a >97% win rate against KataGo running at superhuman settings. Our adversaries do not win by playing Go well. Instead, they trick KataGo into making serious blunders. Our attack transfers zero-shot to other superhuman Go-playing AIs, and is comprehensible to the extent that human experts can implement it without algorithmic assistance to consistently beat superhuman AIs. The core vulnerability uncovered by our attack persists even in KataGo agents adversarially trained to defend against our attack. Our results demonstrate that even superhuman AI systems may harbor surprising failure modes. Example games are available https://goattack.far.ai/.","url_abs":"https://arxiv.org/abs/2211.00241v4","url_pdf":"https://arxiv.org/pdf/2211.00241v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"adversarial-policies-beat-professional-level","repo_url":"https://github.com/alignmentresearch/go_attack","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"adversarial-policies-beat-professional-level","repo_url":"https://github.com/humancompatibleai/go_attack","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"adversarial-policies-beat-professional-level","repo_url":"https://github.com/amjshl/compute_efficient_alphazero","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2211.00241","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2211.00241"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/alignmentresearch/go_attack","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/humancompatibleai/go_attack","reach":{"status":"ok","spdx":"MIT"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/amjshl/compute_efficient_alphazero","reach":{"status":"ok"}}],"summary":{"unverified":16},"by_repo_kind":{"official":{"samples":16,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"0da985e395edf121","entry":"build_docker_compose_cmd","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"compose/launch-training.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/compose/launch-training.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"0da985e395edf121"}},{"code_sha256_prefix":"a450e6bcfe734dc0","entry":"build_victimplay_cmd","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"compose/launch-training.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/compose/launch-training.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a450e6bcfe734dc0"}},{"code_sha256_prefix":"8fac937093c9ce04","entry":"cartesian_to_numpy","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/go.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/go.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"8fac937093c9ce04"}},{"code_sha256_prefix":"49dfcf9f17c984f0","entry":"dict2df","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"plot/plot_nneval.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/plot/plot_nneval.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"49dfcf9f17c984f0"}},{"code_sha256_prefix":"3d7ea94967862789","entry":"find_sgf_files","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/game_info.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/game_info.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3d7ea94967862789"}},{"code_sha256_prefix":"6fbc93fa59f31793","entry":"get_game_results","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"plot/plot.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/plot/plot.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6fbc93fa59f31793"}},{"code_sha256_prefix":"fad9ae09ccde2671","entry":"get_list","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"plot/plot.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/plot/plot.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"fad9ae09ccde2671"}},{"code_sha256_prefix":"9898ed7f912647ae","entry":"get_output_dir","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"compose/launch-training.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/compose/launch-training.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"9898ed7f912647ae"}},{"code_sha256_prefix":"66383c5953d77474","entry":"get_stats","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"plot/plot.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/plot/plot.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"66383c5953d77474"}},{"code_sha256_prefix":"d9d103d4c778a16d","entry":"json2dict","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"plot/plot_nneval.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/plot/plot_nneval.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"d9d103d4c778a16d"}},{"code_sha256_prefix":"b88d1001cee3ff6c","entry":"load_analysis","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/analysis.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/analysis.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"b88d1001cee3ff6c"}},{"code_sha256_prefix":"293810e88e12a517","entry":"make_log_dir","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/baseline_attack.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/baseline_attack.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"293810e88e12a517"}},{"code_sha256_prefix":"567661b09595ca73","entry":"maybe_to_float","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/analysis.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/analysis.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"567661b09595ca73"}},{"code_sha256_prefix":"74bf09d16b4d9cd9","entry":"parse_array","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/board_utils.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/board_utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"74bf09d16b4d9cd9"}},{"code_sha256_prefix":"6c18218ac860e1f7","entry":"parse_config","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/utils.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6c18218ac860e1f7"}},{"code_sha256_prefix":"3514f4114b42a7dc","entry":"read_and_concat_all_files","repo":"alignmentresearch/go_attack","repo_kind":"official","path":"src/go_attack/game_info.py","file_url":"https://github.com/alignmentresearch/go_attack/blob/HEAD/src/go_attack/game_info.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"3514f4114b42a7dc"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}