Claim Verification on PrimeFacts Five Class
42Macro F1 ScoreLlama-3.3-70B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Llama-3.3-70BExtraction Mode=C (open)2026.05 | 42 | 85 | |
| Qwen3-235BExtraction Mode=C (open)2026.05 | 39 | 79 | |
| Llama-3.3-70BExtraction Mode=B (decontext.)2026.05 | 35 | 77 | |
| Qwen3-14BExtraction Mode=C (open)2026.05 | 34 | 74 | |
| Llama-4-ScoutExtraction Mode=C (open)2026.05 | 34 | 78 | |
| Qwen3-32BExtraction Mode=C (open)2026.05 | 33 | 75 | |
| Qwen3-235BExtraction Mode=B (decontext.)2026.05 | 30 | 46 | |
| Qwen3-32BExtraction Mode=B (decontext.)2026.05 | 28 | 64 | |
| Llama-3.3-70BExtraction Mode=A (linked)2026.05 | 28 | 93 | |
| Qwen3-14BExtraction Mode=B (decontext.)2026.05 | 27 | 63 | |
| Qwen3-32BExtraction Mode=A (linked)2026.05 | 27 | 71 | |
| Llama-4-ScoutExtraction Mode=A (linked)2026.05 | 27 | 74 | |
| Llama-4-ScoutExtraction Mode=B (decontext.)2026.05 | 27 | 65 | |
| Qwen3-235BExtraction Mode=A (linked)2026.05 | 26 | 83 | |
| Qwen3-8BExtraction Mode=C (open)2026.05 | 25 | 71 | |
| Qwen3-14BExtraction Mode=A (linked)2026.05 | 23 | 74 | |
| Qwen3-8BExtraction Mode=B (decontext.)2026.05 | 22 | 63 | |
| Qwen3-8BExtraction Mode=A (linked)2026.05 | 21 | 75 | |
| Baseline2026.05 | 10 | — |