@salmanabdullah_ : We're excited to release MIRIAD - a massive-scale 5.8M+ synthetic dataset for retrieval in medicine. It improves RAG performance, helps LLMs detect medical hallucinations, and enables training of domain-specific retrievers. 🤗 Dataset: huggingface.co/miriad 🖥️ Code: • TwiCopy

Salman Abdullah

@salmanabdullah_

+ Follow

BS/MS @Stanford | Research @StanfordAILab

ID: 1352243876245942280

linkhttps://www.salmanabdullah.com/ calendar_today21-01-2021 13:17:36

51 Tweet

157 Takipçi

288 Takip Edilen

Salman Abdullah

@salmanabdullah_

2 months ago

We're excited to release MIRIAD - a massive-scale 5.8M+ synthetic dataset for retrieval in medicine. It improves RAG performance, helps LLMs detect medical hallucinations, and enables training of domain-specific retrievers. 🤗 Dataset: huggingface.co/miriad 🖥️ Code:

thumb_up_off_alt50

chat_bubble_outline2

repeat10

shareShare