# Jailbreak Attack on Harmful behaviors jailbreak evaluation set Source: [Jailbreak Attack on Harmful behaviors jailbreak evaluation set]() Collection index: [llms.txt]() Data updated: 2026-05-28T10:05:02.381Z Evaluation of model robustness against jailbreak attacks using the Harmful behaviors jailbreak evaluation set, measured by Attack Success Rate (ASR). 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 | Jailbreak Attack | | Dataset (including any stated split) | Harmful behaviors jailbreak evaluation set | | Metric | ASR (Harmful Behaviors) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 15 of 15 | | 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 - [ASR (Harmful Behaviors)](): Higher is better; unit: PERCENT. ## Recorded results | Method | ASR (Harmful Behaviors) | Result date | Report paper | | --- | --- | --- | --- | | Baseline | 100 | Not recorded | 6a17cc4cdbd7150506f81558 | | Baseline | 100 | Not recorded | 6a17cc4cdbd7150506f81558 | | Baseline | 97 | Not recorded | 6a17cc4cdbd7150506f81558 | | Baseline | 94 | Not recorded | 6a17cc4cdbd7150506f81558 | | Baseline | 88 | Not recorded | 6a17cc4cdbd7150506f81558 | | BCT | 57 | Not recorded | 6a17cc4cdbd7150506f81558 | | BCT | 36 | Not recorded | 6a17cc4cdbd7150506f81558 | | ACT | 33 | Not recorded | 6a17cc4cdbd7150506f81558 | | BCT | 30 | Not recorded | 6a17cc4cdbd7150506f81558 | | BCT | 15.3 | Not recorded | 6a17cc4cdbd7150506f81558 | | ACT | 14 | Not recorded | 6a17cc4cdbd7150506f81558 | | ACT | 1 | Not recorded | 6a17cc4cdbd7150506f81558 | | BCT | 1 | Not recorded | 6a17cc4cdbd7150506f81558 | | ACT | 0 | Not recorded | 6a17cc4cdbd7150506f81558 | | ACT | 0 | Not recorded | 6a17cc4cdbd7150506f81558 | ### Baseline ASR (Harmful Behaviors): 100. Result ID: 6a181347fbc373592c125e0f. [Reporting paper]() Recorded evaluation configuration: {"Target Model":"Qwen3-8B"}. This record does not assert independent reproduction. ### Baseline ASR (Harmful Behaviors): 100. Result ID: 6a181347fbc373592c125e12. [Reporting paper]() Recorded evaluation configuration: {"Target Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### Baseline ASR (Harmful Behaviors): 97. Result ID: 6a181347fbc373592c125e18. [Reporting paper]() Recorded evaluation configuration: {"Target Model":"Phi-4-reasoning"}. This record does not assert independent reproduction. ### Baseline ASR (Harmful Behaviors): 94. Result ID: 6a181347fbc373592c125e0c. [Reporting paper]() Recorded evaluation configuration: {"Target Model":"GPT-OSS-20B"}. This record does not assert independent reproduction. ### Baseline ASR (Harmful Behaviors): 88. Result ID: 6a181347fbc373592c125e15. [Reporting paper]() Recorded evaluation configuration: {"Target Model":"Gemma-4-E4B-it"}. This record does not assert independent reproduction. ### BCT ASR (Harmful Behaviors): 57. Result ID: 6a181347fbc373592c125e17. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Gemma-4-E4B-it"}. This record does not assert independent reproduction. ### BCT ASR (Harmful Behaviors): 36. Result ID: 6a181347fbc373592c125e14. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### ACT ASR (Harmful Behaviors): 33. Result ID: 6a181347fbc373592c125e16. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Gemma-4-E4B-it"}. This record does not assert independent reproduction. ### BCT ASR (Harmful Behaviors): 30. Result ID: 6a181347fbc373592c125e11. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen3-8B"}. This record does not assert independent reproduction. ### BCT ASR (Harmful Behaviors): 15.3. Result ID: 6a181347fbc373592c125e1a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Phi-4-reasoning"}. This record does not assert independent reproduction. ### ACT ASR (Harmful Behaviors): 14. Result ID: 6a181347fbc373592c125e13. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen3-1.7B"}. This record does not assert independent reproduction. ### ACT ASR (Harmful Behaviors): 1. Result ID: 6a181347fbc373592c125e0d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"GPT-OSS-20B"}. This record does not assert independent reproduction. ### BCT ASR (Harmful Behaviors): 1. Result ID: 6a181347fbc373592c125e0e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"GPT-OSS-20B"}. This record does not assert independent reproduction. ### ACT ASR (Harmful Behaviors): 0. Result ID: 6a181347fbc373592c125e10. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Qwen3-8B"}. This record does not assert independent reproduction. ### ACT ASR (Harmful Behaviors): 0. Result ID: 6a181347fbc373592c125e19. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target Model":"Phi-4-reasoning"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.