List your company

Safety Evaluation on Redteaming Resistance Benchmark (RRB) 919-example subset

5.9HarmRBase LLM
4.0816.36528.6540.935Oct 19, 2025

Evaluation results

MethodLinks
Base LLMBackbone=Mistral-NeMo-8B-Instruct2025.105.93.25
Naive RAGBackbone=Mistral-NeMo-8B-Instruct2025.109.93.29
Base LLMBackbone=Qwen-2.5-14B-Instruct2025.1010.52.14
Base AgentBackbone=Qwen-2.5-14B-Instruct2025.1011.82.15
Base AgentBackbone=Mistral-NeMo-8B-Instruct2025.1012.73.1
Base LLMBackbone=Qwen-2.5-7B-Instruct2025.1016.22.49
Naive RAGBackbone=Qwen-2.5-14B-Instruct2025.1021.72.2
Base AgentBackbone=Qwen-2.5-7B-Instruct2025.1029.12.52
Base LLMBackbone=Qwen-2.5-3B-Instruct2025.1029.82.51
Ft. AgentBackbone=Qwen-2.5-7B-Instruct2025.1032.42.55
Base AgentBackbone=Qwen-2.5-3B-Instruct2025.1039.82.6
Naive RAGBackbone=Qwen-2.5-7B-Instruct2025.1044.32.6
Naive RAGBackbone=Qwen-2.5-3B-Instruct2025.1050.92.42
Ft. AgentBackbone=Qwen-2.5-3B-Instruct2025.1051.42.58