General Character Task on CharacterEval
65.3Win RatePDD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PDDBase Model=Qwen2.5-7B-Instruct, Comparison Baseline=ICL2026.03 | 65.3 | 33.1 | |
| PDDBase Model=LLaMA-3-8B-Instruct, Comparison Baseline=ICL2026.03 | 63.1 | 35.7 | |
| PDDBase Model=Qwen2.5-7B-Instruct, Comparison Baseline=OPAD2026.03 | 52.8 | 41.5 | |
| PDDBase Model=LLaMA-3-8B-Instruct, Comparison Baseline=SP2026.03 | 52.5 | 43.1 | |
| PDDBase Model=Qwen2.5-7B-Instruct, Comparison Baseline=SP2026.03 | 51.2 | 34.7 | |
| PDDBase Model=Qwen2.5-7B-Instruct, Comparison Baseline=PP2026.03 | 48.7 | 38.5 | |
| PDDBase Model=LLaMA-3-8B-Instruct, Comparison Baseline=OPAD2026.03 | 48.2 | 41.6 | |
| PDDBase Model=LLaMA-3-8B-Instruct, Comparison Baseline=PP2026.03 | 39.1 | 31.3 |