Natural Language Inference on ANLI R2
32.62F1 ScoreFLIPPED
Evaluation Results
| Method | Links | |
|---|---|---|
| FLIPPEDModel Size=3B, Evaluation Protocol=Zero-shot2022.10 | 32.62 | |
| FLIPPEDModel Size=11B, Evaluation Protocol=Zero-shot2022.10 | 32.59 | |
| T0Model Size=11B, Evaluation Protocol=Zero-shot2022.10 | 31.25 | |
| DirectModel Size=3B, Evaluation Protocol=Zero-shot2022.10 | 28.23 | |
| T0Model Size=3B, Evaluation Protocol=Zero-shot2022.10 | 23.73 | |
| GPT-3Model Size=175B, Evaluation Protocol=Zero-shot2022.10 | 22.5 | |
| ChannelModel Size=3B, Evaluation Protocol=Zero-shot2022.10 | 21.44 |