Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

35 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Lightweight RL's Razor Reproduction for LLM Reasoning Fine-Tuning

This repository contains a lightweight research-oriented implementation inspired by RL's Razor, focused on comparing supervised fine-tuning and reinforcement learning fine-tuning for mathematical reasoning in small language models.

The project studies whether a simple REINFORCE-style reinforcement learning method can improve reasoning ability while causing less distributional shift from the original base model compared with supervised fine-tuning (SFT).


Research Question

**Does REINFORCE-style post-training minimize representation-level distributional shift and preserve core capabilities better than SFT when optimizing lightweight language models for mathematical reasoning? **

In this project, the comparison is not between different model architectures. Instead, the same base model is adapted using different post-training strategies:

Base Qwen2.5-0.5B-Instruct
├── Base model: no fine-tuning
├── SFT model: supervised fine-tuning on GSM8K
└── RL model: REINFORCE-style fine-tuning with correctness-based rewards

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages