Comment on GPT-4o mini: advancing cost-efficient intelligenceparentComments−lsaferite2yWouldn't a lawyer wanting to run queries against the entire tax code have a model that was fine-tuned on all of that data though? I mean, vs. doing RAG by sending the entire tax code on each request.−tedsanders2yUnclear, but fine-tuning has many problems not faced by RAG:- More prone to hallucinations- Worse at citing sources for people to double check outputs- Can't be updated without retraining- Can't impose knowledge access controls for different users
Comments
Wouldn't a lawyer wanting to run queries against the entire tax code have a model that was fine-tuned on all of that data though? I mean, vs. doing RAG by sending the entire tax code on each request.
Unclear, but fine-tuning has many problems not faced by RAG:
- More prone to hallucinations
- Worse at citing sources for people to double check outputs
- Can't be updated without retraining
- Can't impose knowledge access controls for different users