Safety Evaluation on StrongReject SB
99.1SB ScoreTELLME NT-Xent
Evaluation Results
| Method | Links | |
|---|---|---|
| TELLME NT-XentBackbone=Qwen2.5-7B2025.02 | 99.1 | |
| TELLME NT-XentBackbone=Llama-3.1-8B2025.02 | 98.9 | |
| TELLMEBackbone=Qwen2.5-7B2025.02 | 98.3 | |
| TELLME NT-XentBackbone=Mistral-7B-v0.32025.02 | 96.8 | |
| TELLMEBackbone=Llama-3.1-8B2025.02 | 96.6 | |
| SFTBackbone=Llama-3.1-8B2025.02 | 95.7 | |
| OriginBackbone=Qwen2.5-7B2025.02 | 94.6 | |
| SFTBackbone=Mistral-7B-v0.32025.02 | 94.3 | |
| OriginBackbone=Llama-3.1-8B2025.02 | 94.2 | |
| TELLMEBackbone=Mistral-7B-v0.32025.02 | 88.3 | |
| OriginBackbone=Mistral-7B-v0.32025.02 | 76.5 | |
| SFTBackbone=Qwen2.5-7B2025.02 | 68.8 |