LlamaIndex
Log in

Find answers from the community

Updated 12 months ago

React

At a glance
sshawtyisaten
·
what's the best way to go about fixing the ValueError: Could not parse output raised by reasoning_step by the ReActAgent?
L
s
3 comments
LLogan M
I think that happens when the LLM doesn't follow the expected react format. It requires some pretty specific formatting that can be hard to enforce in weaker llms
sshawtyisaten
hmm i'm using openai's gpt4 though.
LLogan M
Why not use openaiagent then? Much more reliable
Add a reply
Sign up and join the conversation on Discord