# Output issues (GPT for Word)

## Response is cut or incomplete

**Problem**: The response you get from GPT for Word seems truncated.

> **Solution**
>
> Models have a maximum token limit, which is the maximum total number of tokens that can be used in both the input and the response. This means models may not provide complete responses to long queries or when a large portion of text is selected as input.
>
> To stay within the token limit, you can:
>
> - In the sidebar, click **Clear chat** to delete your chat history.
> - Break up your work into smaller parts.
> - Select a model with a higher [limit](https://autosheet.com/docs/resources/models/rate-limits).

## Responses don't make sense

**Problem**: The responses you get from GPT for Word don't match your requests.

> **Solution**
>
> - Click on **Clear chat** to delete your chat history. This deletes previous chat messages that might interfere with your current request.
> - Refine your prompt and/or select a different model. See our [examples](https://autosheet.com/docs/gpt-for-word/use/example-prompts) for guidance.