Factual Reasoning on HaluEval General
30Baseline WinsDomain-Grounded Tiered Retrieval
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Domain-Grounded Tiered RetrievalN=1502026.03 | 30 | 75 | 45 | 50 | 21.2 | 78.8 | |
| Baseline (Llama 3.1 8B)N=1502026.03 | 30 | — | — | — | — | — |
Loading the SOTA2 catalog…