Conversational Recommendation on Home & Kitchen
49.5Success RateD2D
Evaluation Results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| D2DModel=o4-mini2026.06 | 49.5 | 16.67 | 41.56 | 49.5 | 0.5 | 2.81 | 424 | 92,418 | 1,192 | 259,694 | |
| D2DModel=gpt-4.1-mini2026.06 | 47 | 14.89 | 40.9 | 51.5 | 1 | 3.54 | 265 | 91,541 | 936 | 323,385 | |
| Full-LLMModel=gpt-4.1-mini2026.06 | 40.5 | 10.72 | 21.18 | 53 | 6 | 4.75 | 833 | 85,402 | 3,962 | 406,085 | |
| Full-LLMModel=o4-mini2026.06 | 24.5 | 6.34 | 10.47 | 74 | 1 | 2.56 | 1,386 | 84,840 | 3,540 | 216,768 | |
| ZS RecommenderModel=gpt-4.1-mini2026.06 | 24 | 5.98 | 11.02 | 69 | 6.5 | 1.43 | 528 | 83,987 | 755 | 120,101 | |
| ZS RecommenderModel=o4-mini2026.06 | 12.5 | 3.41 | 5.86 | 86.5 | 0.5 | 2.13 | 658 | 89,194 | 1,402 | 189,983 | |
| Embedding RetrievalModel=N/A2026.06 | 6 | 1.3 | 8.79 | — | — | — | — | — | — | — | |
| BM25 RetrievalModel=N/A2026.06 | 5 | 1.16 | 7.32 | — | — | — | — | — | — | — |