Personalized workflow prediction on DRFLOW (mixed split)
94.38FactualityDRFA
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DRFABackbone=GPT-5.22026.06 | 94.38 | 90.04 | 41.36 | 55.67 | 65.86 | 47.96 | 53.67 | |
| DRBABackbone=Claude-Opus-4.52026.06 | 86.78 | 93.1 | 23.59 | 39.56 | 63.7 | 37.86 | 44.52 | |
| DRBABackbone=GPT-5.22026.06 | 86.41 | 86.52 | 24.58 | 42.61 | 58.13 | 37.3 | 44.36 | |
| DRBABackbone=Average2026.06 | 83.4 | 87.19 | 19.3 | 35.74 | 54.53 | 32.44 | 38.27 | |
| DRFABackbone=Average2026.06 | 82.89 | 86.77 | 33.74 | 52.08 | 58.7 | 37.95 | 43.91 | |
| DRFABackbone=Claude-Opus-4.52026.06 | 82.54 | 89.68 | 46.72 | 58.51 | 68.35 | 47.43 | 53.19 | |
| DRFABackbone=DeepSeek-v3.22026.06 | 82.3 | 82.03 | 28.03 | 51.83 | 51.35 | 32.32 | 38.34 | |
| DRBABackbone=DeepSeek-v3.22026.06 | 81.32 | 83.13 | 16.12 | 27.26 | 48.55 | 30.36 | 34.04 | |
| DRBABackbone=Gemini-3.1-pro2026.06 | 79.1 | 86 | 12.92 | 33.54 | 47.75 | 24.24 | 30.18 | |
| DRFABackbone=Gemini-3.1-pro2026.06 | 72.36 | 85.32 | 18.86 | 42.32 | 49.26 | 24.07 | 30.42 |