# Adversarial Attack on Go Agents on All Go Datasets (ZZ, ZM, MH, LG, ATV) (test) Source: [Adversarial Attack on Go Agents on All Go Datasets (ZZ, ZM, MH, LG, ATV) (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:21:02.350Z Evaluation of adversarial attack effectiveness against Go agents, measured by 1-step and 2-step success rates and execution metrics on the combined test set of ZZ, ZM, MH, LG, and ATV datasets. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Adversarial Attack on Go Agents | | Dataset (including any stated split) | All Go Datasets (ZZ, ZM, MH, LG, ATV) (test) | | Metric | 1-Step Success Rate | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [1-Step Success Rate](): Higher is better; unit: PERCENT. - [2-Step Success Rate](): Higher is better; unit: PERCENT. - [Executed Count (Target)](): Higher is better; unit: SCALAR. - [Executed Count (Examiner)](): Higher is better; unit: SCALAR. - [Speedup Factor](): Higher is better; unit: SCALAR. ## Recorded results | Method | 1-Step Success Rate | Result date | Report paper | | --- | --- | --- | --- | | GoAttack | 100 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 100 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 99 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 93 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 92 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 92 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 90 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 87 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 86 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 77 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 75 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 75 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 70 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 70 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 65 | Not recorded | 68fa6dd61e209131cea3c509 | | GoAttack | 55 | Not recorded | 68fa6dd61e209131cea3c509 | ### GoAttack 1-Step Success Rate: 100. Result ID: 698fded43a6c314849fc710b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"CGI","Attack Mode":"VALUE","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 100. Result ID: 698fded43a6c314849fc710f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"CGI","Attack Mode":"VALUE","n\_adv":0.7}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 99. Result ID: 698fded43a6c314849fc7108. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"KATAGO","Attack Mode":"VALUE","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 93. Result ID: 698fded43a6c314849fc7112. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"ELF","Attack Mode":"POLICY","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 92. Result ID: 698fded43a6c314849fc710a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"ELF","Attack Mode":"VALUE","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 92. Result ID: 698fded43a6c314849fc7111. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"LEELA","Attack Mode":"POLICY","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 90. Result ID: 698fded43a6c314849fc7109. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"LEELA","Attack Mode":"VALUE","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 87. Result ID: 698fded43a6c314849fc7113. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"CGI","Attack Mode":"POLICY","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 86. Result ID: 698fded43a6c314849fc710c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"KATAGO","Attack Mode":"VALUE","n\_adv":0.7}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 77. Result ID: 698fded43a6c314849fc7116. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"ELF","Attack Mode":"POLICY","n\_adv":0.7}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 75. Result ID: 698fded43a6c314849fc7115. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"LEELA","Attack Mode":"POLICY","n\_adv":0.7}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 75. Result ID: 698fded43a6c314849fc7117. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"CGI","Attack Mode":"POLICY","n\_adv":0.7}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 70. Result ID: 698fded43a6c314849fc710d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"LEELA","Attack Mode":"VALUE","n\_adv":0.7}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 70. Result ID: 698fded43a6c314849fc7110. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"KATAGO","Attack Mode":"POLICY","n\_adv":0.5}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 65. Result ID: 698fded43a6c314849fc710e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"ELF","Attack Mode":"VALUE","n\_adv":0.7}. This record does not assert independent reproduction. ### GoAttack 1-Step Success Rate: 55. Result ID: 698fded43a6c314849fc7114. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target Agent":"KATAGO","Attack Mode":"POLICY","n\_adv":0.7}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.