Synergizing Reasoning and Acting in Language Models
Alternatives Considered
No response
Additional Context
No response
Originally created by @xlxxcc on GitHub (Apr 10, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/12682
### Check Existing Issues
- [x] I have searched the existing issues and discussions.
### Problem Description
Have you considered supporting ReAct? Paper link: https://arxiv.org/abs/2210.03629
### Desired Solution you'd like
Synergizing Reasoning and Acting in Language Models
### Alternatives Considered
_No response_
### Additional Context
_No response_
it's a 3 years old code, i think that it's been integrated & overpassed,
by the way: They chose a really bad short name!!!
<!-- gh-comment-id:2792217862 -->
@rgaricano commented on GitHub (Apr 10, 2025):
it's a 3 years old code, i think that it's been integrated & overpassed,
by the way: They chose a really bad short name!!!
What do you mean by that? If you want to use ReAct as merely a new prompt technique, then there would be no issues.
Suppose we want to generate a product description for a new smartphone. We can use ReAct prompting to generate a description that highlights the phone's features and benefits.
Reason: Generate a list of the phone's features, such as its camera resolution, battery life, and storage capacity.
Act: Use the generated features to create a product description that showcases the phone's benefits and unique selling points.
However, I struggle to think of a solution for supporting Reason+Act without an API or tools involved. Maybe you can achieve this more advance usage using LiteLLM instead.
<!-- gh-comment-id:2793953013 -->
@bgeneto commented on GitHub (Apr 10, 2025):
What do you mean by that? If you want to use ReAct as merely a new prompt technique, then there would be no issues.
```
Suppose we want to generate a product description for a new smartphone. We can use ReAct prompting to generate a description that highlights the phone's features and benefits.
Reason: Generate a list of the phone's features, such as its camera resolution, battery life, and storage capacity.
Act: Use the generated features to create a product description that showcases the phone's benefits and unique selling points.
```
However, I struggle to think of a solution for supporting Reason+Act without an API or tools involved. Maybe you can achieve this more advance usage using LiteLLM instead.
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 @xlxxcc on GitHub (Apr 10, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/12682
Check Existing Issues
Problem Description
Have you considered supporting ReAct? Paper link: https://arxiv.org/abs/2210.03629
Desired Solution you'd like
Synergizing Reasoning and Acting in Language Models
Alternatives Considered
No response
Additional Context
No response
@rgaricano commented on GitHub (Apr 10, 2025):
it's a 3 years old code, i think that it's been integrated & overpassed,
by the way: They chose a really bad short name!!!
@bgeneto commented on GitHub (Apr 10, 2025):
What do you mean by that? If you want to use ReAct as merely a new prompt technique, then there would be no issues.
However, I struggle to think of a solution for supporting Reason+Act without an API or tools involved. Maybe you can achieve this more advance usage using LiteLLM instead.
@tjbck commented on GitHub (Apr 10, 2025):
It's already supported as "native" tool calling.