Typesetting Optimization on PaperFit-Bench 1.0 (test)
1Compile ScorePaperFit
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PaperFit2026.05 | 1 | 1 | 3.3907 | 89.5 | 4.579 | 80.5 | |
| VisualMRfeedback_signal=rendered page images, turn_mode=multi-round2026.05 | 0.975 | 0.975 | 2.8006 | 65 | 4.5789 | 54.87 | |
| VisualSTfeedback_signal=rendered page images, turn_mode=single-turn2026.05 | 0.625 | 0.625 | 1.8741 | 29.5 | 2.7681 | 45.6 | |
| TextMRfeedback_signal=text, turn_mode=multi-round2026.05 | 0.61 | 0.61 | 2.1601 | 42.5 | 2.7433 | 62.3 | |
| TextSTfeedback_signal=text, turn_mode=single-turn2026.05 | 0.585 | 0.585 | 1.8522 | 28 | 2.5738 | 45.3 | |
| Perturbedstatus=baseline2026.05 | 0.58 | 0.82 | 1.8275 | 0 | 3.6344 | 37.5 | |
| RuleLogfeedback_signal=compile logs2026.05 | 0.52 | 0.76 | 2.1838 | 38 | 3.3401 | 44.44 |