czg1225/VeriThinker

[NeurIPS 2025] VeriThinker: Learning to Verify Makes Reasoning Model Efficient

28
/ 100
Experimental

This project helps make large language models (LLMs) more efficient at complex reasoning tasks, like solving math problems, by training them to verify their own steps. You provide a question, and the model generates a shorter, more accurate step-by-step solution compared to standard methods. This is ideal for researchers and practitioners working with advanced AI models who need to reduce computational costs while maintaining or improving accuracy in problem-solving.

No commits in the last 6 months.

Use this if you need to make your large language models perform complex reasoning with fewer computational steps and higher accuracy.

Not ideal if you are looking for an off-the-shelf application for end-users, as this is a tool for developers working with LLMs.

AI model efficiency Large Language Models Reasoning optimization Computational linguistics research AI model training
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 8 / 25
Maturity 15 / 25
Community 3 / 25

How are scores calculated?

Stars

65

Forks

1

Language

Python

License

MIT

Last pushed

Sep 27, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/czg1225/VeriThinker"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.