OpenJudge: A Unified Framework for Holistic Evaluation and Quality Rewards

🤖 Automate Bing Searches 🔍, Quizzes 🧪, Polls 📝, & more for Bing Rewards. 💸

对ChatGLM直接使用RLHF提升或降低目标输出概率|Modify ChatGLM output with only RLHF

👤 Multi-Armed Bandit Algorithms Library (MAB) 👮

This repository contains the code for SFT, RLHF, and DPO, designed for vision-based LLMs, including the LLaVA models and the LLaMA-3.2-vision models.

Scalable and extensible reinforcement learning for LM agents.

An advanced desktop automation tool for Microsoft Rewards. It performs Bing searches and collects Daily Sets using mathematically driven, human-like input simulation (W3C Actions, Bezier curves, and smart scrolling). Built with Python/Selenium and packaged as an executable Windows app for a seamless, plug-and-play experience.

official implementation of ICLR'2025 paper: Rethinking Bradley-Terry Models in Preference-based Reward Modeling: Foundations, Theory, and Alternatives

The official CLI for the Lumen Protocol & Local Prompt Generation.

[ICLR 2026] Unleashing Scientific Reasoning for Bio-experimental Protocol Generation via Structured Component-based Reward Mechanism

During our participation in the Internship Exchange Program, my friend and I collaborated with the guidance of our esteemed supervisor from NTHU.

chatglm_rlhf_finetuning

realize the reinforcement learning training for gpt2 llama bloom and so on llm model

Worksheet and Utilities for AWS DeepRacer – one of the most exciting ways of building strong skills in reinforcement learning and through a hands-on approach. This repository offers: 1) Functionally-rich and flexible reward function 2) Utilities with Jupiter notes for Racing Line calculation and visualisation of track 3) Scripts to parse RoboMak…

用RLHF可选LoRA对LLaMA和MOSS进行训练|Training LLaMA or MOSS with RLHF [LoRA]

Creating an environment to quickly train a variety of Deep Reinforcement Learning algorithms on Street Fighter 2 using tournaments between learning agents

Value & Policy Iteration for the frozenlake environment of OpenAI

Deep Reinforcement Learning with LTL goals.

Script to automagically add channel currency for users that use channel point rewards. Potentially deprecated due to twitch api changes.