19 lines
480 B
Markdown
19 lines
480 B
Markdown
|
---
|
||
|
|
license: mit
|
||
|
|
library_name: transformers
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
tags:
|
||
|
|
- trl
|
||
|
|
- grpo
|
||
|
|
- conflict-env
|
||
|
|
- reasoning
|
||
|
|
---
|
||
|
|
|
||
|
|
# ConflictEnv Final Reasoning Model
|
||
|
|
|
||
|
|
This is the final fine-tuned model for the **ConflictEnv** executive assistant task.
|
||
|
|
It has been trained using **GRPO** to handle complex scheduling conflicts with a focus on reasoning-first behavior.
|
||
|
|
|
||
|
|
## Usage
|
||
|
|
Start prompts with `Scenario: ... Details: ...` and expect a `<thought>` block followed by a JSON action.
|