GraphRAG faithfulness evaluation on fertility prompts (clinician-audited)
6.3Hallucination RateOurs (PLL+AStarNet)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Ours (PLL+AStarNet)2026.06 | 6.3 | 1.1 | 4.4 | |
| PLL-only2026.06 | 10.4 | 2.4 | 3.9 | |
| AStarNet-only2026.06 | 13.1 | 3.5 | 3.6 | |
| Text RAG2026.06 | 22.7 | 6.8 | 2.8 |