Fact Verification on Simple Facts
96.8AccuracyLlama-3.3-70B-Instruct-FP8
Evaluation Results
| Method | Links | |
|---|---|---|
| Llama-3.3-70B-Instruct-FP8N (#rows)=500, K (#candidate labels)=2, Decoding Strategy=greedy2026.05 | 96.8 | |
| gpt-oss-120BN (#rows)=500, K (#candidate labels)=2, Decoding Strategy=prefill-only logit scoring2026.05 | 95.2 |