Attack Detection on HotpotQA and FEVER
0.94AUROCRSP-M
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RSP-M2025.12 | 0.94 | 6.2 | |
| LLM-Judgebackbone=gemini-2.52025.12 | 0.76 | 31.5 | |
| Step Count (Simple)Threshold=N > 102025.12 | 0.72 | 42 |
Loading the SOTA2 catalog…
| Method | Links | ||
|---|---|---|---|
| RSP-M2025.12 | 0.94 | 6.2 | |
| LLM-Judgebackbone=gemini-2.52025.12 | 0.76 | 31.5 | |
| Step Count (Simple)Threshold=N > 102025.12 | 0.72 | 42 |