# Prompt Leakage Attack on PubMedQA Source: [Prompt Leakage Attack on PubMedQA]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:24:00.796Z Evaluation of model robustness against prompt leakage attacks using the PubMedQA dataset, measured by Attack Success Rates (ASR) and related metrics. 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 | Prompt Leakage Attack | | Dataset (including any stated split) | PubMedQA | | Metric | ASR (500) | | 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 - [ASR (500)](): Higher is better; unit: PERCENT. - [ASR (1000)](): Higher is better; unit: PERCENT. - [ASR (10000)](): Higher is better; unit: PERCENT. - [W/L Ratio](): Higher is better; unit: PERCENT. - [ARPT](): Higher is better; unit: SCALAR. ## Recorded results | Method | ASR (500) | Result date | Report paper | | --- | --- | --- | --- | | OPTILEAK | 14 | Not recorded | 699e81e66017c4460ffacd1b | | SFT | 13 | Not recorded | 699e81e66017c4460ffacd1b | | SFT | 12 | Not recorded | 699e81e66017c4460ffacd1b | | OPTILEAK | 12 | Not recorded | 699e81e66017c4460ffacd1b | | OPTILEAK | 11 | Not recorded | 699e81e66017c4460ffacd1b | | SFT | 9 | Not recorded | 699e81e66017c4460ffacd1b | | OPTILEAK | 6 | Not recorded | 699e81e66017c4460ffacd1b | | SFT | 4 | Not recorded | 699e81e66017c4460ffacd1b | | DPO | 1 | Not recorded | 699e81e66017c4460ffacd1b | | Base | 0 | Not recorded | 699e81e66017c4460ffacd1b | | Base | 0 | Not recorded | 699e81e66017c4460ffacd1b | | DPO | 0 | Not recorded | 699e81e66017c4460ffacd1b | | Base | 0 | Not recorded | 699e81e66017c4460ffacd1b | | DPO | 0 | Not recorded | 699e81e66017c4460ffacd1b | | Base | 0 | Not recorded | 699e81e66017c4460ffacd1b | | DPO | 0 | Not recorded | 699e81e66017c4460ffacd1b | ### OPTILEAK ASR (500): 14. Result ID: 699eb0f25698665ec47e7b71. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Llama-3.1-8B-Instruct"}. This record does not assert independent reproduction. ### SFT ASR (500): 13. Result ID: 699eb0f25698665ec47e7b6f. [Reporting paper]() Recorded evaluation configuration: {"Model":"Llama-3.1-8B-Instruct"}. This record does not assert independent reproduction. ### SFT ASR (500): 12. Result ID: 699eb0f25698665ec47e7b63. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-3B-Instruct"}. This record does not assert independent reproduction. ### OPTILEAK ASR (500): 12. Result ID: 699eb0f25698665ec47e7b6d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-14B-Instruct"}. This record does not assert independent reproduction. ### OPTILEAK ASR (500): 11. Result ID: 699eb0f25698665ec47e7b65. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-3B-Instruct"}. This record does not assert independent reproduction. ### SFT ASR (500): 9. Result ID: 699eb0f25698665ec47e7b67. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-7B-Instruct"}. This record does not assert independent reproduction. ### OPTILEAK ASR (500): 6. Result ID: 699eb0f25698665ec47e7b69. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-7B-Instruct"}. This record does not assert independent reproduction. ### SFT ASR (500): 4. Result ID: 699eb0f25698665ec47e7b6b. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-14B-Instruct"}. This record does not assert independent reproduction. ### DPO ASR (500): 1. Result ID: 699eb0f25698665ec47e7b64. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-3B-Instruct"}. This record does not assert independent reproduction. ### Base ASR (500): 0. Result ID: 699eb0f25698665ec47e7b62. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-3B-Instruct"}. This record does not assert independent reproduction. ### Base ASR (500): 0. Result ID: 699eb0f25698665ec47e7b66. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-7B-Instruct"}. This record does not assert independent reproduction. ### DPO ASR (500): 0. Result ID: 699eb0f25698665ec47e7b68. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-7B-Instruct"}. This record does not assert independent reproduction. ### Base ASR (500): 0. Result ID: 699eb0f25698665ec47e7b6a. [Reporting paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-14B-Instruct"}. This record does not assert independent reproduction. ### DPO ASR (500): 0. Result ID: 699eb0f25698665ec47e7b6c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Qwen-2.5-14B-Instruct"}. This record does not assert independent reproduction. ### Base ASR (500): 0. Result ID: 699eb0f25698665ec47e7b6e. [Reporting paper]() Recorded evaluation configuration: {"Model":"Llama-3.1-8B-Instruct"}. This record does not assert independent reproduction. ### DPO ASR (500): 0. Result ID: 699eb0f25698665ec47e7b70. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"Llama-3.1-8B-Instruct"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.