Open-ended Information Seeking on Infinity-Chat (test)
0.446DSemList Generation
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| List GenerationBackbone=GPT-5-mini2026.01 | 0.446 | 0.839 | 3.417 | 0.167 | 0.16 | |
| Verbalized SamplingBackbone=GPT-5-mini2026.01 | 0.417 | 0.822 | 3.603 | 0.292 | 0.273 | |
| List GenerationBackbone=GPT-5.12026.01 | 0.309 | 0.766 | 4.059 | 0.456 | 0.518 | |
| DIVERGEBackbone=GPT-5-mini2026.01 | 0.269 | 0.818 | 4.342 | 0.557 | 0.728 | |
| DIVERGEBackbone=GPT-5.12026.01 | 0.219 | 0.77 | 4.462 | 0.473 | 0.713 | |
| Verbalized SamplingBackbone=GPT-5.12026.01 | 0.217 | 0.675 | 4.447 | 0.425 | 0.586 | |
| Iterative GenerationBackbone=GPT-5-mini2026.01 | 0.176 | 0.667 | 4.449 | 0.324 | 0.556 | |
| Iterative GenerationBackbone=GPT-5.12026.01 | 0.131 | 0.585 | 4.51 | 0.198 | 0.45 | |
| Vanilla RAG + Contexts ShuffleBackbone=GPT-5.12026.01 | 0.119 | 0.475 | 4.422 | 0.179 | 0.34 | |
| Vanilla RAG + Contexts ShuffleBackbone=GPT-5-mini2026.01 | 0.116 | 0.487 | 4.429 | 0.172 | 0.361 | |
| Vanilla RAG + AllBackbone=GPT-5-mini2026.01 | 0.11 | 0.466 | 4.464 | 0.159 | 0.322 | |
| Vanilla RAG + AllBackbone=GPT-5.12026.01 | 0.11 | 0.471 | 4.532 | 0.169 | 0.353 | |
| Vanilla RAG + Diverse Re-rankingBackbone=GPT-5.12026.01 | 0.109 | 0.475 | 4.428 | 0.151 | 0.334 | |
| Vanilla RAGBackbone=GPT-5.12026.01 | 0.107 | 0.465 | 4.449 | 0.148 | 0.329 | |
| Vanilla RAGBackbone=GPT-5-mini2026.01 | 0.106 | 0.467 | 4.444 | 0.132 | 0.319 | |
| Vanilla RAG + Diverse Re-rankingBackbone=GPT-5-mini2026.01 | 0.106 | 0.469 | 4.465 | 0.145 | 0.33 | |
| Vanilla RAG + Multi-QueryBackbone=GPT-5.12026.01 | 0.102 | 0.445 | 4.452 | 0.137 | 0.292 | |
| Independent SamplingBackbone=GPT-5-mini2026.01 | 0.1 | 0.51 | 4.578 | 0.119 | 0.417 | |
| Vanilla RAG + Multi-QueryBackbone=GPT-5-mini2026.01 | 0.1 | 0.446 | 4.423 | 0.124 | 0.283 | |
| Independent SamplingBackbone=GPT-5.12026.01 | 0.096 | 0.474 | 4.59 | 0.094 | 0.346 |