Directness on C3 news community level post-cutoff (test)
61Win PercentageMRC
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MRCHierarchy level=LF, Evaluation model=GPT-3.5-turbo2026.03 | 61 | 28 | 11 | |
| MRCHierarchy level=L1, Evaluation model=GPT-3.5-turbo2026.03 | 59 | 28 | 13 | |
| M2hCHierarchy level=LF, Evaluation model=GPT-3.5-turbo2026.03 | 58 | 31 | 11 | |
| M2hCHierarchy level=L1, Evaluation model=GPT-3.5-turbo2026.03 | 56 | 31 | 13 | |
| RkHHierarchy level=LF, Evaluation model=GPT-3.5-turbo2026.03 | 48 | 42 | 10 | |
| RkHHierarchy level=L1, Evaluation model=GPT-3.5-turbo2026.03 | 45 | 45 | 10 |