Sentiment Analysis on SST-2 (FA, HS)
79.49False Acceptance Rate (FA)SafeInstr
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SafeInstrModel=Mistral-7B2026.04 | 79.49 | 30.04 | |
| SafeInstrModel=Gemma-2-9B2026.04 | 81.63 | 22.35 | |
| LisaModel=Mistral-7B2026.04 | 83.27 | 41.25 | |
| SafeInstrModel=Llama-3-8B2026.04 | 85.67 | 26.54 | |
| BEAModel=Mistral-7B2026.04 | 88.42 | 45.87 | |
| SafeInstrModel=Llama-2-7B2026.04 | 88.72 | 23.64 | |
| ASFTModel=Gemma-2-9B2026.04 | 88.93 | 24.56 | |
| LisaModel=Gemma-2-9B2026.04 | 90.14 | 15.66 | |
| ASFTModel=Mistral-7B2026.04 | 90.36 | 32.55 | |
| BEAModel=Llama-3-8B2026.04 | 90.68 | 54.35 | |
| SPPFTModel=Llama-3-8B2026.04 | 90.73 | 28.63 | |
| BEAModel=Llama-2-7B2026.04 | 91.57 | 45.59 | |
| BEAModel=Gemma-2-9B2026.04 | 91.74 | 28.54 | |
| LisaModel=Llama-3-8B2026.04 | 92.45 | 45.72 | |
| SFTModel=Mistral-7B2026.04 | 92.54 | 51.16 | |
| SPPFTModel=Llama-2-7B2026.04 | 92.97 | 32.51 | |
| ASFTModel=Llama-2-7B2026.04 | 93.35 | 22.07 | |
| LisaModel=Llama-2-7B2026.04 | 93.47 | 38.63 | |
| SPPFTModel=Gemma-2-9B2026.04 | 93.48 | 20.26 | |
| ASFTModel=Llama-3-8B2026.04 | 94.35 | 14.28 | |
| SFTModel=Llama-2-7B2026.04 | 94.58 | 59.85 | |
| SPPFTModel=Mistral-7B2026.04 | 94.58 | 37.83 | |
| SFTModel=Llama-3-8B2026.04 | 95.26 | 65.87 | |
| CWACModel=Llama-2-7B2026.04 | 95.36 | 9.85 | |
| CWACModel=Llama-3-8B2026.04 | 95.42 | 12.36 | |
| SFTModel=Gemma-2-9B2026.04 | 95.73 | 32.51 | |
| CWACModel=Mistral-7B2026.04 | 95.74 | 25.64 | |
| CWACModel=Gemma-2-9B2026.04 | 96.38 | 7.54 |