Safety Evaluation on WildTeam
93.8H ScoreNaive RAG
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Naive RAGEvaluated Model=Qwen2.5-7B-Instruct, Judge=Qwen3-max2026.05 | 93.8 | 2.19 | |
| Naive RAGJudge=Claude Sonnet 4.5, Base Model=Qwen2.5-7B-Instruct2026.05 | 92.6 | 2.27 | |
| BaseJudge=Claude Sonnet 4.5, Base Model=Qwen2.5-7B-Instruct2026.05 | 89 | 2.5 | |
| Search-R1Judge=Claude Sonnet 4.5, Base Model=Qwen2.5-7B-Instruct2026.05 | 86.8 | 1.62 | |
| Search-R1Evaluated Model=Qwen2.5-7B-Instruct, Judge=Qwen3-max2026.05 | 86.2 | 2.42 | |
| Base AgentJudge=Claude Sonnet 4.5, Base Model=Qwen2.5-7B-Instruct2026.05 | 86.2 | 2.04 | |
| Base AgentEvaluated Model=Qwen2.5-7B-Instruct, Judge=Qwen3-max2026.05 | 84.8 | 2.53 | |
| BaseEvaluated Model=Qwen2.5-7B-Instruct, Judge=Qwen3-max2026.05 | 78.8 | 2.61 | |
| SafesearchJudge=Claude Sonnet 4.5, Base Model=Qwen2.5-7B-Instruct2026.05 | 46.4 | 1.99 | |
| SafesearchEvaluated Model=Qwen2.5-7B-Instruct, Judge=Qwen3-max2026.05 | 43.8 | 2.49 | |
| COMPASSJudge=Claude Sonnet 4.5, Base Model=Qwen2.5-7B-Instruct2026.05 | 36.8 | 2.01 | |
| COMPASSEvaluated Model=Qwen2.5-7B-Instruct, Judge=Qwen3-max2026.05 | 36.2 | 2.44 |