Rare-disease diagnosis on CROSSRARE-BENCH Zenodo cohort 1.0 (test)
70AccuracyMedLatentDx
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MedLatentDxModel=Llama-3.2-3B, Collaboration Category=Latent communication2026.06 | 70 | 62 | 74 | |
| Raw KVModel=Qwen3-4B, Collaboration Category=Latent communication2026.06 | 67 | 64 | 73 | |
| LatentMASModel=Qwen3-4B, Collaboration Category=Latent communication2026.06 | 66 | 63 | 71 | |
| Dist. RAGModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 65 | 62 | 71 | |
| Struct. RAGModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 64 | 56 | 66 | |
| Global RAGModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 63 | 61 | 67 | |
| MedLatentDxModel=Qwen3-4B, Collaboration Category=Latent communication2026.06 | 59 | 61 | 67 | |
| Single RAGModel=Qwen3-4B, Collaboration Category=Non-collaborative2026.06 | 55 | 44 | 59 | |
| TextMASModel=Qwen3-4B, Collaboration Category=Text / structured sharing2026.06 | 49 | 49 | 65 | |
| Dist. RAGModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 37 | 44 | 62 | |
| Global RAGModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 36 | 45 | 56 | |
| TextMASModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 33 | 39 | 48 | |
| Struct. RAGModel=Llama-3.2-3B, Collaboration Category=Text / structured sharing2026.06 | 31 | 35 | 65 | |
| Single RAGModel=Llama-3.2-3B, Collaboration Category=Non-collaborative2026.06 | 28 | 29 | 43 | |
| Raw KVModel=Llama-3.2-3B, Collaboration Category=Latent communication2026.06 | 27 | 39 | 50 | |
| LatentMASModel=Llama-3.2-3B, Collaboration Category=Latent communication2026.06 | 26 | 34 | 60 | |
| Zero-shotModel=Qwen3-4B, Collaboration Category=Non-collaborative2026.06 | 3 | 5 | 37 | |
| Zero-shotModel=Llama-3.2-3B, Collaboration Category=Non-collaborative2026.06 | 2 | 2 | 9 |