search

gpt-engineer-org / gpt-engineer

Platform to experiment with the AI Software Engineer. Terminal based. NOTE: Very different from https://gptengineer.app
MIT License
52.12k stars 6.79k forks source link

No asking for clarification in qa agent #213

Closed hellin-eth closed 1 year ago

hellin-eth commented 1 year ago

Maybe it is because i dont have openai api for the gpt4 model, but even changing the identity agent 'qa' to: Then you will pick one clarifying question, and wait for an answer from the user. Even if there is no need of clarification, wait for the answer from the user.

there is no clarification asking. Maybe I am doing something wrong?

AntonOsika commented 1 year ago

Did you append --steps-config clarify?

LeonardoIBO commented 1 year ago

When i answer the clarification questions for some reason it just outputs a "No" back again and continues with the process. It's kinda funny but probably not expected behavior. Any idea why this is happening?

hellin-eth commented 1 year ago

No, i haven't! I thought it was a feature, not a execution parameter. Thanks a lot, this project is funny as f***!!

I'll close this issue, as a problem between keyboard and chair.

flrivera commented 9 months ago

Did you append --steps-config clarify?

where do we append this, when running gpte projects/example gpt-3.5-turbo-0613 ?

ATheorell commented 9 months ago

@flrivera

try gpte projects/example gpt-3.5-turbo-0613 -c

BUT, sometimes the LLM (in particular 3.5) decides that the instructions are clear enough and doesn't ask questions.