Elizabeth H.

Welcome!

So nice to meet you in person, Matt! Thank you for coming :)

Hi Ourahou O., are you working in python or typescript? I can link some resources. I'll also DM you with a link to sign up for my evals office hours

Truong H. right now the best way is to use the _dumps method, like you mentioned. We will look into adding a utility for more graceful conversion, though!

Kyle J. this is also likely related to how the LLM response is parsed when not using tool-calling. You should be able to set use_function_calling_if_available=True when you run llm_classify with your evaluator. What model are you using?

Nico G., have you tried increasing the timeout from the default?

In the typescript library, we delegate to the AI SDK from Vercel, but there isn't one clear "winner" on the python side.

You're the first person we've heard of using this llm tool. We decided to start with adaptors to the 3 most common LLM clients we see people using: openai, langchain, and litellm. Between those 3, pretty much every model is supported. We intentionally avoided strict dependency on any one package, such that users can choose which SDK they prefer to use.