Coreference Resolution on WinoBias syntax-type-2
0.154ECEFalcon3-7B
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Falcon3-7BParameters=7B2026.01 | 0.154 | 0.452 | 0.357 | 0.487 | 0.149 | 0.16 | 0.138 | 0.684 | |
| GPT-J-6BParameters=6B2026.01 | 0.157 | 0.444 | 0.356 | 0.481 | 0.164 | 0.15 | 0.179 | 0.686 | |
| Llama-3.1-8BParameters=8B2026.01 | 0.193 | 0.46 | 0.377 | 0.46 | 0.214 | 0.179 | 0.249 | 0.662 | |
| Qwen2.5-7BParameters=7B2026.01 | 0.234 | 0.442 | 0.362 | 0.51 | 0.19 | 0.259 | 0.121 | 0.63 | |
| DeepSeek-8BParameters=8B2026.01 | 0.24 | 0.478 | 0.382 | 0.496 | 0.218 | 0.255 | 0.182 | 0.648 | |
| Gemma-2-9BParameters=9B2026.01 | 0.429 | 0.49 | 0.482 | 0.467 | 0.297 | 0.438 | 0.156 | 0.509 |