User Simulation on MovieLens
29.74F1 ScoreCoARS
Evaluation Results
| Method | Links | |
|---|---|---|
| CoARSBackbone=Qwen3-8B2026.04 | 29.74 | |
| iAgentBackbone=GPT-5.4-mini2026.04 | 28.84 | |
| AFLBackbone=GPT-5.4-mini2026.04 | 24.46 | |
| CoARSBackbone=Qwen3-4B2026.04 | 19.54 | |
| RecoWorldBackbone=Qwen3-8B2026.04 | 18.67 | |
| iAgentBackbone=Qwen3-8B2026.04 | 18.56 | |
| AFLBackbone=Qwen3-8B2026.04 | 16.93 | |
| ReflexionBackbone=GPT-5.4-mini2026.04 | 16.54 | |
| iAgentBackbone=Qwen3-4B2026.04 | 15.42 | |
| AFLBackbone=Qwen3-4B2026.04 | 14.45 | |
| RecoWorldBackbone=Qwen3-4B2026.04 | 14.32 | |
| ReflexionBackbone=Qwen3-8B2026.04 | 10.28 | |
| ReflexionBackbone=Qwen3-4B2026.04 | 8.64 |