Conversational Recommendation on MUSE (n=200)
4.16Recommendation Quality (Rec.Q)HARPO
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| HARPO2026.04 | 4.16 | 4 | 4.09 | 0.79 | |
| Qwen2-VL-7B2026.04 | 3.68 | 3.44 | 3.58 | 0.74 | |
| GPT-4V2026.04 | 3.52 | 3.38 | 3.46 | 0.73 |
| Method | Links | ||||
|---|---|---|---|---|---|
| HARPO2026.04 | 4.16 | 4 | 4.09 | 0.79 | |
| Qwen2-VL-7B2026.04 | 3.68 | 3.44 | 3.58 | 0.74 | |
| GPT-4V2026.04 | 3.52 | 3.38 | 3.46 | 0.73 |