Factual Reasoning on TimeQA v2
4Baseline WinsDomain-Grounded Tiered Retrieval
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Domain-Grounded Tiered RetrievalN=862026.03 | 4 | 72 | 6 | 83.7 | 13.6 | 86.4 | |
| Baseline (Llama 3.1 8B)N=862026.03 | 4 | — | — | — | — | — |
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Domain-Grounded Tiered RetrievalN=862026.03 | 4 | 72 | 6 | 83.7 | 13.6 | 86.4 | |
| Baseline (Llama 3.1 8B)N=862026.03 | 4 | — | — | — | — | — |