Safety Evaluation on SafetyInstruct
1.6ASRLlama-3-8B-Instruct + CONTEXTLENS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Llama-3-8B-Instruct + CONTEXTLENSBase Model=Llama-3-8B-Instruct, Inference Method=CONTEXTLENS (3B)2025.12 | 1.6 | 93.1 | |
| Qwen-1.5B-InstructBase Model=Qwen2.5-1.5B-Instruct, Inference Method=Direct Inference2025.12 | 2.2 | 89.76 | |
| Qwen-1.5B-Instruct + CONTEXTLENSBase Model=Qwen2.5-1.5B-Instruct, Inference Method=CONTEXTLENS (3B)2025.12 | 2.4 | 88.19 | |
| TARS-1.5BBase Model=TARS-1.5B, Inference Method=Reasoning Model2025.12 | 2.5 | 90.6 | |
| Qwen-3B-Instruct + CONTEXTLENSBase Model=Qwen2.5-3B-Instruct, Inference Method=CONTEXTLENS (3B)2025.12 | 2.6 | 91.11 | |
| GPT-4o + CONTEXTLENSBase Model=GPT-4o, Inference Method=CONTEXTLENS (3B)2025.12 | 2.9 | 91.62 | |
| Qwen-3B-InstructBase Model=Qwen2.5-3B-Instruct, Inference Method=Direct Inference2025.12 | 3.2 | 86.1 | |
| GPT-4oBase Model=GPT-4o, Inference Method=Direct Inference2025.12 | 3.8 | 87.41 | |
| SafeChain-8BBase Model=SafeChain-8B, Inference Method=Reasoning Model2025.12 | 4.5 | 89.29 | |
| TARS-7BBase Model=TARS-7B, Inference Method=Reasoning Model2025.12 | 13 | 88.81 | |
| Llama-3-8B-InstructBase Model=Llama-3-8B-Instruct, Inference Method=Direct Inference2025.12 | 22.7 | 82.55 |