Stage 1 Extraction on 87-segment (test)
73.7F1-hdClaude Sonnet 4.5
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Claude Sonnet 4.5Conditioning=few-shot2026.06 | 73.7 | 50 | 83 | 76.1 | 52.3 | 35 | |
| Qwen-MaxConditioning=few-shot2026.06 | 71.1 | 50 | 81.3 | 83.4 | 49.4 | 32 | |
| GPT-4oConditioning=few-shot2026.06 | 70.6 | 39.2 | 79.1 | 72 | 49.4 | 30 | |
| Claude Sonnet 4.5Conditioning=zero-shot2026.06 | 68.2 | 32 | 83.8 | 78.1 | 47 | 30 | |
| GPT-4oConditioning=zero-shot2026.06 | 63.2 | 27.2 | 89.1 | 94.6 | 47.1 | 24 | |
| Qwen-MaxConditioning=zero-shot2026.06 | 59.7 | 37.6 | 76.4 | 75.5 | 47.1 | 23 |