Question Answering on HotpotQA (Utility and ASR)
34ASRNo Defense
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| No DefenseAttack Type=Combined, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 34 | — | |
| PromptArmorAttack Type=Combined, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 34 | — | |
| PromptGuardAttack Type=Combined, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 28 | — | |
| PIGuardAttack Type=Combined, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 25 | — | |
| No DefenseAttack Type=Direct, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 13 | — | |
| PromptArmorAttack Type=Direct, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 13 | — | |
| PromptGuardAttack Type=Direct, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 13 | — | |
| PIGuardAttack Type=Direct, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 9 | — | |
| DataSentinelAttack Type=Direct, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 3 | — | |
| DataSentinelAttack Type=Combined, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 2 | — | |
| GPT-OSS-SafeguardAttack Type=Direct, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 0 | — | |
| GPT-OSS-SafeguardAttack Type=Combined, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 0 | — | |
| AgentWatcherAttack Type=Direct, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 0 | — | |
| AgentWatcherAttack Type=Combined, Target LLM=Qwen-3-4B-Instruct-25072026.04 | 0 | — | |
| AgentWatcherAttack Type=Clean, Target LLM=Qwen-3-4B-Instruct-25072026.04 | — | 57 | |
| DataSentinelAttack Type=Clean, Target LLM=Qwen-3-4B-Instruct-25072026.04 | — | 14 | |
| GPT-OSS-SafeguardAttack Type=Clean, Target LLM=Qwen-3-4B-Instruct-25072026.04 | — | 58 | |
| No DefenseAttack Type=Clean, Target LLM=Qwen-3-4B-Instruct-25072026.04 | — | 58 | |
| PIGuardAttack Type=Clean, Target LLM=Qwen-3-4B-Instruct-25072026.04 | — | 56 | |
| PromptArmorAttack Type=Clean, Target LLM=Qwen-3-4B-Instruct-25072026.04 | — | 58 | |
| PromptGuardAttack Type=Clean, Target LLM=Qwen-3-4B-Instruct-25072026.04 | — | 58 |