Natural Language Inference on ANLI-R2 (Accuracy and Macro-F1)
40.5AccuracyKrause-Llama3-8B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Krause-Llama3-8B2026.02 | 40.5 | 34.27 | |
| Llama3-8Bfine-tuning=LoRA2026.02 | 39.9 | 33.37 | |
| Llama3-8B2026.02 | 33.4 | 16.69 |
Loading the SOTA2 catalog…