Quick note here:
I've been working on a domain-specific GPT for parsing FDA/clinical/genomic data. The hardest part so far is making it handle messy trial docs + RNAseq outputs without hallucination.
Curious if others have tried similar-or thoughts on good eval?
Quick note here:
I've been working on a domain-specific GPT for parsing FDA/clinical/genomic data. The hardest part so far is making it handle messy trial docs + RNAseq outputs without hallucination.
Curious if others have tried similar-or thoughts on good eval?