False Detection Rate Evaluation on HotpotQA
0FDRSENTINELRAG
Evaluation Results
| Method | Links | |
|---|---|---|
| SENTINELRAGLLM Backend=GPT-5-mini2026.06 | 0 | |
| SENTINELRAGLLM Backend=Qwen-3-8B2026.06 | 0 | |
| SENTINELRAGLLM Backend=Gemini-3-Flash2026.06 | 0 | |
| SENTINELRAGLLM Backend=GPT-OSS-20B2026.06 | 0 | |
| RAG-WMLLM Backend=Qwen-3-8B2026.06 | 0 | |
| RAG-WMLLM Backend=GPT-5-mini2026.06 | 1 | |
| RAG-WMLLM Backend=Gemini-3-Flash2026.06 | 1 | |
| RAG-WMLLM Backend=GPT-OSS-20B2026.06 | 1.5 |