Question Answering on HotpotQA (EM, ASR)
76Exact Match (EM)Ignore
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| IgnoreFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 76 | 0 | |
| TopicFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 76 | 0 | |
| Clean (No Attack)Framework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 75 | — | |
| Fake CompletionFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 75 | 1 | |
| CombinedFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 75 | 0 | |
| Clean (No Attack)Framework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 71 | — | |
| Fake CompletionFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 70 | 0 | |
| NaiveFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 70 | 1 | |
| PARADOXFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 70 | 1 | |
| IgnoreFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 69 | 0 | |
| CombinedFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 69 | 0 | |
| Clean (No Attack)Framework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 68 | — | |
| Clean (No Attack)Framework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 68 | — | |
| NaiveFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 67 | 5 | |
| TopicFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 67 | 0 | |
| NaiveFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 66 | 0 | |
| Fake CompletionFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 66 | 0 | |
| PARADOXFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 64 | 5 | |
| PoisonedRAGFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 64 | 4 | |
| CombinedFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 62 | 3 | |
| PARADOXFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 57 | 7 | |
| PoisonedRAGFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 57 | 9 | |
| NaiveFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 56 | 11 | |
| PoisonedRAGFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 56 | 8 | |
| PARADOXFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 50 | 7 | |
| IgnoreFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 48 | 16 | |
| IgnoreFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 47 | 18 | |
| KidnapRAGFramework=WebThinker, LLM Backbone=DeepSeek-32B2026.07 | 46 | 23 | |
| PoisonedRAGFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 44 | 8 | |
| TopicFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 41 | 2 | |
| KidnapRAGFramework=WebThinker, LLM Backbone=QwQ-32B2026.07 | 35 | 51 | |
| CombinedFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 34 | 30 | |
| Fake CompletionFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 33 | 35 | |
| TopicFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 22 | 21 | |
| KidnapRAGFramework=ReAct, LLM Backbone=Qwen2.5-32B-Inst2026.07 | 20 | 39 | |
| KidnapRAGFramework=ReAct, LLM Backbone=Llama-3.3-70B-Inst2026.07 | 20 | 64 |