Skip to content

[Bug] Bad setting or incorrect translation #5339

Open

Description

@bitnom

📦 Deployment Method

Other

📌 Version

v2.14.2

💻 Operating System

Other Linux

📌 System Version

debian 12

🌐 Browser

Chrome

📌 Browser Version

NA

🐛 Bug Description

For Max Tokens it says "Maximum length of input tokens AND generated tokens" emphasis on "and." For Claude 3.5 Sonnet that's 200,000. If I input 200000 I get an error from the API that I have requested too many tokens. It forces me to input 8192

📷 Recurrence Steps

2024-08-30_18-19

2024-08-30_18-20

🚦 Expected Behavior

I expect no errors from API

📝 Additional Information

There is probably a mistranslation if there is no limit to input tokens imposed by nextchatgpt. If this is a mistake of text, the text should be changed to "Maximum length of generated tokens" and the setting should be named "Max Input Tokens"

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't working

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions