Sentiment Analysis on SST2 (Accuracy, Consistency Rate, Overconfidence Rate)
94.27AccuracyDGAO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DGAOModel=Qwen-2.5-14B2026.05 | 94.27 | 98.91 | 7.41 | |
| SFTModel=Qwen-2.5-14B2026.05 | 94.15 | 97.05 | 7.52 | |
| DGAOModel=Qwen-2.5-3B2026.05 | 93.53 | 99.35 | 6.76 | |
| SFTModel=Qwen-2.5-3B2026.05 | 93.48 | 97.33 | 6.91 | |
| PAFTModel=Qwen-2.5-3B2026.05 | 92.45 | 99.28 | 7.18 | |
| PAFTModel=Qwen-2.5-14B2026.05 | 92.23 | 98.88 | 7.75 | |
| DGAOModel=Llama-3.2-11B2026.05 | 91.31 | 99.27 | 8.41 | |
| DGAOModel=Llama-3.2-3B2026.05 | 90.64 | 99.14 | 8.8 | |
| SFTModel=Llama-3.2-11B2026.05 | 90.23 | 99.18 | 9.87 | |
| DGAOModel=Gemma-3-4B2026.05 | 90.11 | 99.03 | 9.48 | |
| SFTModel=Gemma-3-4B2026.05 | 89.98 | 98.95 | 9.27 | |
| SFTModel=Llama-3.2-3B2026.05 | 89.47 | 99.08 | 10.26 | |
| PAFTModel=Gemma-3-4B2026.05 | 88.64 | 99.66 | 11.25 | |
| PAFTModel=Llama-3.2-11B2026.05 | 88.47 | 99.24 | 11.43 | |
| PAFTModel=Llama-3.2-3B2026.05 | 87.62 | 99.08 | 11.87 |