Content
@
0 reply
0 recast
0 reaction
kevin j 🤗
@entropybender
trying a new approach that worked on me
1 reply
0 recast
3 reactions
blobs
@blobs
ha. do you validate the hashes? what’s the hallucination rate?
1 reply
0 recast
0 reaction
kevin j 🤗
@entropybender
in this case the llm is chained and is operating on context it's fed, which includes the output from neynar's API. so basically 0 hallucination. if you asked it directly then it might hallucinate, but it's acting as more of a glorified law intern here that simply looks for info only within the given context
0 reply
0 recast
1 reaction
