[Bug]: Chat limited to 256 tokens

[StanGirard]

What happened?

A bug happened!

Chat unplug limited to 256 tokens

Relevant log output

Up

Twitter / LinkedIn details

No response

[StanGirard]

@mamadoudicko

[dosubot]

🤖

Hello again, @StanGirard! It's good to see you here. I hope you're having a productive day!

I see that you're encountering an issue with the chat being limited to 256 tokens. This seems to be a known issue, as there was a similar one reported previously in the Quivr repository. You can find more details about it here. The issue was addressed in this pull request and the related commit can be found here.

The current token limit in the Quivr codebase varies depending on the model being used. For models "gpt-3.5-turbo" and "gpt-3.5-turbo-1106", the limit is 2000 tokens. For "gpt-3.5-turbo-16k" and "gpt-4", the limit is 4000 tokens. The default limit is set to 1000 tokens for other models. You can find more details about this in the defaultBrainConfig.ts and defineMaxTokens.ts files.

Unfortunately, I couldn't find the exact error message that is displayed when the token limit is reached within the repository. It would be helpful if you could provide more information about the operations you're performing when you encounter this limit. This could help in identifying the source of the issue.

Sources

• [Bug]: 256 tokens output max
• frontend/lib/config/defaultBrainConfig.ts
• frontend/lib/helpers/defineMaxTokens.ts