Strengths3
- Provides a clear demonstration of sycophancy reduction via DPO fine-tuning
- Includes evaluation code for standard sycophancy benchmarks (SycoBench)
- Offers an interactive Gradio app for comparing model behaviors
Tutorial / Demo · Jupyter Notebook
This repository provides tools to evaluate and demonstrate a Qwen2.5-1.5B model fine-tuned with DPO to reduce sycophancy, including baseline and post-training evaluation notebooks, a dataset cleaning script, and a Gradio app for interactive comparison.
Merit describes the code at this commit. Later pushes count once the owner rescans. How merit is scored
Each field's share of the project's AI work, as scoring counts it. It ranks in the field it is most about.
Read from the project's code at its scored commit.
The owner keeps this project's limitations private.