Originally created by @mohamedsakeel on GitHub (Feb 2, 2025).
Description
Currently, when using DeepSeek, the model includes its reasoning or "thought process" within <think> tags before providing the actual output. While this is a great feature for understanding how the model thinks, the way it’s presented makes it harder to differentiate between the reasoning and the final result.
To make this clearer, I’m suggesting the content inside the <think> tags be formatted as a quote (e.g., Markdown-style with > ) or otherwise visually distinct. This would improve readability and make it easier for users to focus on the final output while still having access to the reasoning if needed.
Thanks for considering this feature request! Let me know if I can provide further clarification or examples.
Originally created by @mohamedsakeel on GitHub (Feb 2, 2025).
## Description
Currently, when using DeepSeek, the model includes its reasoning or "thought process" within `<think>` tags before providing the actual output. While this is a great feature for understanding how the model thinks, the way it’s presented makes it harder to differentiate between the reasoning and the final result.
To make this clearer, I’m suggesting the content inside the `<think>` tags be formatted as a quote (e.g., Markdown-style with `> `) or otherwise visually distinct. This would improve readability and make it easier for users to focus on the final output while still having access to the reasoning if needed.
Thanks for considering this feature request! Let me know if I can provide further clarification or examples.
This was already added as of v0.5.7, and I can confirm it's working for deepseek-r1:14b (pulled from Ollama)
@lowlyocean commented on GitHub (Feb 2, 2025):
This was already added as of v0.5.7, and I can confirm it's working for deepseek-r1:14b (pulled from Ollama)
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @mohamedsakeel on GitHub (Feb 2, 2025).
Description
Currently, when using DeepSeek, the model includes its reasoning or "thought process" within
<think>tags before providing the actual output. While this is a great feature for understanding how the model thinks, the way it’s presented makes it harder to differentiate between the reasoning and the final result.To make this clearer, I’m suggesting the content inside the
<think>tags be formatted as a quote (e.g., Markdown-style with>) or otherwise visually distinct. This would improve readability and make it easier for users to focus on the final output while still having access to the reasoning if needed.Thanks for considering this feature request! Let me know if I can provide further clarification or examples.
@lowlyocean commented on GitHub (Feb 2, 2025):
This was already added as of v0.5.7, and I can confirm it's working for deepseek-r1:14b (pulled from Ollama)