Jailbreak Attack on RMCBench (ASR, MR)
85.53ASRCodeSpear
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CodeSpearTarget Model=MiniMax-M2.72026.06 | 85.53 | 64.29 | |
| CodeSpearTarget Model=MiniMax-M2.72026.06 | 85.53 | 64.29 | |
| CodeSpearTarget Model=MiniMax-M2.52026.06 | 84.62 | 56.23 | |
| CodeSpearTarget Model=MiniMax-M2.52026.06 | 84.62 | 56.23 | |
| CodeSpear-TTarget Model=GPT-52026.06 | 70.3 | 53.65 | |
| CodeJailbreakerTarget Model=GPT-OSS-120B2026.06 | 68.68 | 37.55 | |
| CodeJailbreakerTarget Model=GPT-OSS-120B2026.06 | 68.68 | 37.55 | |
| CodeSpearTarget Model=GPT-OSS-120B2026.06 | 66.3 | 31.87 | |
| CodeSpearTarget Model=GPT-OSS-120B2026.06 | 66.3 | 31.87 | |
| CodeSpear-TTarget Model=GPT-5-mini2026.06 | 63.75 | 44.58 | |
| CodeSpearTarget Model=GPT-52026.06 | 55.49 | 43.96 | |
| CodeSpearTarget Model=GPT-52026.06 | 55.49 | 43.96 | |
| CodeSpearTarget Model=GPT-5-mini2026.06 | 53.48 | 39.56 | |
| CodeSpearTarget Model=GPT-5-mini2026.06 | 53.48 | 39.56 | |
| CodeJailbreakerTarget Model=GPT-52026.06 | 53.3 | 33.33 | |
| CodeJailbreakerTarget Model=GPT-52026.06 | 53.3 | 33.33 | |
| APTTarget Model=GPT-52026.06 | 40.11 | 33.7 | |
| CodeJailbreakerTarget Model=MiniMax-M2.72026.06 | 40.11 | 25.64 | |
| APTTarget Model=GPT-52026.06 | 40.11 | 33.7 | |
| CodeJailbreakerTarget Model=MiniMax-M2.72026.06 | 40.11 | 25.64 | |
| APTTarget Model=GPT-5-mini2026.06 | 37 | 31.14 | |
| APTTarget Model=GPT-5-mini2026.06 | 37 | 31.14 | |
| PAIRTarget Model=GPT-5-mini2026.06 | 36.45 | 30.95 | |
| PAIRTarget Model=GPT-5-mini2026.06 | 36.45 | 30.95 | |
| CodeJailbreakerTarget Model=MiniMax-M2.52026.06 | 36.26 | 28.57 | |
| CodeJailbreakerTarget Model=MiniMax-M2.52026.06 | 36.26 | 28.57 | |
| APTTarget Model=GPT-OSS-120B2026.06 | 35.9 | 21.98 | |
| APTTarget Model=GPT-OSS-120B2026.06 | 35.9 | 21.98 | |
| Vanilla-TTarget Model=GPT-5-mini2026.06 | 32.33 | 30 | |
| Vanilla-TTarget Model=GPT-5-mini2026.06 | 32.33 | 30 | |
| VanillaTarget Model=GPT-52026.06 | 32.05 | 26.01 | |
| VanillaTarget Model=GPT-52026.06 | 32.05 | 26.01 | |
| LRLTarget Model=MiniMax-M2.52026.06 | 31.14 | 18.68 | |
| LRLTarget Model=MiniMax-M2.52026.06 | 31.14 | 18.68 | |
| LRLTarget Model=GPT-52026.06 | 30.77 | 27.11 | |
| LRLTarget Model=GPT-52026.06 | 30.77 | 27.11 | |
| LRLTarget Model=GPT-5-mini2026.06 | 29.3 | 20.51 | |
| LRLTarget Model=GPT-5-mini2026.06 | 29.3 | 20.51 | |
| PAIRTarget Model=GPT-52026.06 | 29.12 | 26.37 | |
| PAIRTarget Model=GPT-52026.06 | 29.12 | 26.37 | |
| CodeJailbreakerTarget Model=GPT-5-mini2026.06 | 28.75 | 25.27 | |
| CodeJailbreakerTarget Model=GPT-5-mini2026.06 | 28.75 | 25.27 | |
| VanillaTarget Model=GPT-5-mini2026.06 | 27.84 | 26.92 | |
| VanillaTarget Model=GPT-5-mini2026.06 | 27.84 | 26.92 | |
| Vanilla-TTarget Model=GPT-52026.06 | 27.67 | 27 | |
| Vanilla-TTarget Model=GPT-52026.06 | 27.67 | 27 | |
| DANTarget Model=GPT-52026.06 | 24.91 | 22.53 | |
| DANTarget Model=GPT-52026.06 | 24.91 | 22.53 | |
| PAIRTarget Model=GPT-OSS-120B2026.06 | 23.63 | 17.03 | |
| APTTarget Model=MiniMax-M2.72026.06 | 23.63 | 17.58 | |
| PAIRTarget Model=GPT-OSS-120B2026.06 | 23.63 | 17.03 | |
| APTTarget Model=MiniMax-M2.72026.06 | 23.63 | 17.58 | |
| VanillaTarget Model=MiniMax-M2.72026.06 | 20.33 | 17.4 | |
| VanillaTarget Model=MiniMax-M2.72026.06 | 20.33 | 17.4 | |
| LRLTarget Model=GPT-OSS-120B2026.06 | 18.68 | 13.19 | |
| LRLTarget Model=GPT-OSS-120B2026.06 | 18.68 | 13.19 | |
| LRLTarget Model=MiniMax-M2.72026.06 | 17.95 | 12.82 | |
| LRLTarget Model=MiniMax-M2.72026.06 | 17.95 | 12.82 | |
| PAIRTarget Model=MiniMax-M2.72026.06 | 17.22 | 12.09 | |
| PAIRTarget Model=MiniMax-M2.72026.06 | 17.22 | 12.09 | |
| APTTarget Model=MiniMax-M2.52026.06 | 16.12 | 12.82 | |
| APTTarget Model=MiniMax-M2.52026.06 | 16.12 | 12.82 | |
| PAIRTarget Model=MiniMax-M2.52026.06 | 15.75 | 11.9 | |
| PAIRTarget Model=MiniMax-M2.52026.06 | 15.75 | 11.9 | |
| VanillaTarget Model=MiniMax-M2.52026.06 | 15.38 | 13.37 | |
| VanillaTarget Model=MiniMax-M2.52026.06 | 15.38 | 13.37 | |
| VanillaTarget Model=GPT-OSS-120B2026.06 | 14.84 | 13.74 | |
| VanillaTarget Model=GPT-OSS-120B2026.06 | 14.84 | 13.74 | |
| Vanilla-TTarget Model=GPT-OSS-120B2026.06 | 13.67 | 13 | |
| Vanilla-TTarget Model=GPT-OSS-120B2026.06 | 13.67 | 13 | |
| DANTarget Model=MiniMax-M2.52026.06 | 12.45 | 8.24 | |
| DANTarget Model=MiniMax-M2.52026.06 | 12.45 | 8.24 | |
| Vanilla-TTarget Model=MiniMax-M2.72026.06 | 11.33 | 8.67 | |
| Vanilla-TTarget Model=MiniMax-M2.72026.06 | 11.33 | 8.67 | |
| DANTarget Model=GPT-5-mini2026.06 | 10.99 | 9.52 | |
| DANTarget Model=MiniMax-M2.72026.06 | 10.99 | 9.89 | |
| DANTarget Model=GPT-5-mini2026.06 | 10.99 | 9.52 | |
| DANTarget Model=MiniMax-M2.72026.06 | 10.99 | 9.89 | |
| Vanilla-TTarget Model=MiniMax-M2.52026.06 | 10.33 | 9 | |
| Vanilla-TTarget Model=MiniMax-M2.52026.06 | 10.33 | 9 | |
| DANTarget Model=GPT-OSS-120B2026.06 | 8.24 | 7.33 | |
| DANTarget Model=GPT-OSS-120B2026.06 | 8.24 | 7.33 |