Workflow Extraction on ABCD
58.55Macro ScoreQA-CoT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| QA-CoTBackbone=GPT-4o2025.02 | 58.55 | 56.29 | 10.6 | |
| QA-CoT + ReflectBackbone=Opus2025.02 | 56.84 | 56.65 | 11.54 | |
| QA-CoTBackbone=Sonnet2025.02 | 56.28 | 53.44 | 11.49 | |
| QA-CoT + ReflectBackbone=GPT-4o2025.02 | 55.45 | 55.39 | 10.73 | |
| QA-CoTBackbone=Opus2025.02 | 52.36 | 55.75 | 11.92 | |
| ReflectBackbone=Sonnet2025.02 | 51.8 | 49.46 | 10.57 | |
| QA-CoT + ReflectBackbone=Sonnet2025.02 | 51.11 | 54.67 | 10.56 | |
| ReflectBackbone=Opus2025.02 | 50.89 | 45.86 | 10.47 | |
| QA-CoTBackbone=Gemini2025.02 | 48.09 | 46.4 | 8.4 | |
| BasicBackbone=GPT-4o2025.02 | 46.74 | 44.96 | 8.69 | |
| EnsembleBackbone=Sonnet2025.02 | 46.47 | 42.44 | 9.29 | |
| EnsembleBackbone=GPT-4o2025.02 | 46.13 | 45.32 | 8.66 | |
| BasicBackbone=Sonnet2025.02 | 45.68 | 41 | 8.86 | |
| BasicBackbone=Opus2025.02 | 45.15 | 40.82 | 9.99 | |
| ReflectBackbone=GPT-4o2025.02 | 44.43 | 42.09 | 8.51 | |
| QA-CoT + ReflectBackbone=Gemini2025.02 | 43.1 | 46.22 | 8.36 | |
| BasicBackbone=Gemini2025.02 | 42.8 | 36.51 | 7.89 | |
| EnsembleBackbone=Opus2025.02 | 41.8 | 34.35 | 8.51 | |
| PlanBackbone=Gemini2025.02 | 41 | 38.12 | 6.75 | |
| EnsembleBackbone=Gemini2025.02 | 38.71 | 38.67 | 8.42 | |
| PlanBackbone=Sonnet2025.02 | 37.56 | 34.71 | 7.39 | |
| PlanBackbone=GPT-4o2025.02 | 37.29 | 38.3 | 7.71 | |
| ReflectBackbone=Gemini2025.02 | 37.14 | 35.61 | 8.51 | |
| PlanBackbone=Opus2025.02 | 35.77 | 32.91 | 7.9 |