Text Generation on BookSum
26.5F1 ScorePrompt-R1
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Prompt-R1Backbone=GPT-4o-mini2025.11 | 26.5 | 60.87 | |
| GRPOBackbone=Qwen3-4B2025.11 | 25.88 | 54.21 | |
| TextGradOptimization Framework=APO, Backbone=GPT-4o-mini2025.11 | 25.79 | 34.95 | |
| SFTBackbone=Qwen3-4B2025.11 | 25.47 | 54.55 | |
| BaselineBackbone=GPT-4o-mini2025.11 | 20.45 | 57.01 | |
| BaselineBackbone=Qwen3-4B2025.11 | 19.99 | 45.19 | |
| CoT ReasoningBackbone=Qwen3-4B2025.11 | 16.68 | 35.02 | |
| OPROOptimization Framework=APO, Backbone=GPT-4o-mini2025.11 | 16.18 | 23.22 | |
| CoT ReasoningBackbone=GPT-4o-mini2025.11 | 11.58 | 56.51 | |
| GEPAOptimization Framework=APO, Backbone=GPT-4o-mini2025.11 | 0 | 0 |