User Simulation on LastFM
31.45F1 ScoreCoARS
Evaluation Results
| Method | Links | |
|---|---|---|
| CoARSBackbone=Qwen3-8B2026.04 | 31.45 | |
| iAgentBackbone=GPT-5.4-mini2026.04 | 26.53 | |
| AFLBackbone=GPT-5.4-mini2026.04 | 24.82 | |
| CoARSBackbone=Qwen3-4B2026.04 | 21.64 | |
| iAgentBackbone=Qwen3-8B2026.04 | 15.72 | |
| RecoWorldBackbone=Qwen3-8B2026.04 | 15.13 | |
| AFLBackbone=Qwen3-8B2026.04 | 14.1 | |
| iAgentBackbone=Qwen3-4B2026.04 | 13.56 | |
| RecoWorldBackbone=Qwen3-4B2026.04 | 12.56 | |
| ReflexionBackbone=GPT-5.4-mini2026.04 | 11.82 | |
| AFLBackbone=Qwen3-4B2026.04 | 11.46 | |
| ReflexionBackbone=Qwen3-8B2026.04 | 8.96 | |
| ReflexionBackbone=Qwen3-4B2026.04 | 6.42 |