I have searched the existing issues and discussions.
Problem Description
(1) I just pitched an idea to my LLM.
(2) Then, I regenerated the answer providing more details (that is, I prompted "try again" but I missclicked "more Details" and it still just got rolling without the details provided).
I repeated step (2) multiple times and recognized: It just made sense of the initial answer and the remarks provided in my last refinement.
Now I want to know which of my eight Answers were a result of which refinement.
Desired Solution you'd like
The easiest way to aid my issue would - mostlikely and from the top of my head - be to simply show refinements of the prompt within the LLM actions done to provide my answer.
Between the "Title of the Model in Use" and the "Reply" it provided, that is.
So as a task-entry the model considered (light-grey) before answering.
Alternatives Considered
looking into log-files.
Additional Context
I was refining an idea.
While providing more context to regenerate the answer the TextInput-Field was very inconvenient.
Its immediate result still differs from a new chat with the refinement.
I am trying to learn your tool.
I am unaware what refinements or retry imply/conclude.
Thank you for your software.
Example
User:
Give me all default vim bindings in a list. Don’t skip any.
Elaborate as necessary but keep it at max. six words per entry.
Skip obvious bindings for an experienced vimmer.
Response: ...list...
Refinement: Put captions of modes and transform it into a (markdown-)table. > Add Details > hit Regenerate.
Result: Not indicated what the refinement has been.
Originally created by @thchha on GitHub (Sep 11, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/17349
### Check Existing Issues
- [X] I have searched the existing issues and discussions.
### Problem Description
(1) I just pitched an idea to my LLM.
(2) Then, I regenerated the answer providing more details (that is, I prompted "try again" but I missclicked "more Details" and it still just got rolling without the </empty> details provided).
I repeated step (2) multiple times and recognized: It just made sense of the initial answer and the remarks provided in my last refinement.
Now I want to know which of my eight Answers were a result of which refinement.
### Desired Solution you'd like
The easiest way to aid my issue would - mostlikely and from the top of my head - be to simply show refinements of the prompt within the LLM actions done to provide my answer.
<img width="433" height="88" alt="Image" src="https://github.com/user-attachments/assets/de5cedd6-637d-44c4-9937-83a59110ca8e" />
Between the "Title of the Model in Use" and the "Reply" it provided, that is.
So as a task-entry the model considered (light-grey) before answering.
### Alternatives Considered
looking into log-files.
### Additional Context
I was refining an idea.
While providing more context to regenerate the answer the TextInput-Field was _very_ inconvenient.
Its immediate result still differs from a new chat with the refinement.
> I am trying to learn your tool.
> I am unaware what refinements or retry imply/conclude.
Thank you for your software.
#### Example
User:
```
Give me all default vim bindings in a list. Don’t skip any.
Elaborate as necessary but keep it at max. six words per entry.
Skip obvious bindings for an experienced vimmer.
```
Response: `...list...`
Refinement: `Put captions of modes and transform it into a (markdown-)table.` > `Add Details` > hit `Regenerate`.
Result: Not indicated what the refinement has been.
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 @thchha on GitHub (Sep 11, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/17349
Check Existing Issues
Problem Description
(1) I just pitched an idea to my LLM.
(2) Then, I regenerated the answer providing more details (that is, I prompted "try again" but I missclicked "more Details" and it still just got rolling without the details provided).
I repeated step (2) multiple times and recognized: It just made sense of the initial answer and the remarks provided in my last refinement.
Now I want to know which of my eight Answers were a result of which refinement.
Desired Solution you'd like
The easiest way to aid my issue would - mostlikely and from the top of my head - be to simply show refinements of the prompt within the LLM actions done to provide my answer.
Between the "Title of the Model in Use" and the "Reply" it provided, that is.
So as a task-entry the model considered (light-grey) before answering.
Alternatives Considered
looking into log-files.
Additional Context
I was refining an idea.
While providing more context to regenerate the answer the TextInput-Field was very inconvenient.
Its immediate result still differs from a new chat with the refinement.
Thank you for your software.
Example
User:
Response:
...list...Refinement:
Put captions of modes and transform it into a (markdown-)table.>Add Details> hitRegenerate.Result: Not indicated what the refinement has been.