Asap7772 / understanding-rlhfLinks
Learning from preferences is a common paradigm for fine-tuning language models. Yet, many algorithmic design decisions come into play. Our new work finds that approaches employing on-policy sampling or negative gradients outperform offline, maximum likelihood objectives.
☆29Updated last year
Alternatives and similar repositories for understanding-rlhf
Users that are interested in understanding-rlhf are comparing it to the libraries listed below
Sorting:
- Dateset Reset Policy Optimization☆30Updated last year
- Code and data used in the paper: "Training on Incorrect Synthetic Data via RL Scales LLM Math Reasoning Eight-Fold"☆30Updated last year
- This is code for most of the experiments in the paper Understanding the Effects of RLHF on LLM Generalisation and Diversity☆43Updated last year
- Directional Preference Alignment☆57Updated 9 months ago
- official implementation of ICLR'2025 paper: Rethinking Bradley-Terry Models in Preference-based Reward Modeling: Foundations, Theory, and…☆62Updated 2 months ago
- Code for Paper (Policy Optimization in RLHF: The Impact of Out-of-preference Data)☆28Updated last year
- Reinforcing General Reasoning without Verifiers☆60Updated last week
- Advantage Leftover Lunch Reinforcement Learning (A-LoL RL): Improving Language Models with Advantage-based Offline Policy Gradients☆26Updated 9 months ago
- Code for the arXiv preprint "The Unreasonable Effectiveness of Easy Training Data"☆48Updated last year
- ☆16Updated 7 months ago
- NeurIPS 2024 tutorial on LLM Inference☆45Updated 6 months ago
- Reinforcement Learning via Regressing Relative Rewards☆34Updated 6 months ago
- Easy-to-Hard Generalization: Scalable Alignment Beyond Human Supervision☆120Updated 9 months ago
- ☆53Updated last week
- The repository contains code for Adaptive Data Optimization☆25Updated 6 months ago
- Official repository of paper "RNNs Are Not Transformers (Yet): The Key Bottleneck on In-context Retrieval"☆27Updated last year
- Self-Supervised Alignment with Mutual Information☆19Updated last year
- ☆32Updated 5 months ago
- Code for "Echo Chamber: RL Post-training Amplifies Behaviors Learned in Pretraining"☆18Updated 2 months ago
- Minimal RLHF implementation built on top of minGPT.☆29Updated 11 months ago
- Code and Configs for Asynchronous RLHF: Faster and More Efficient RL for Language Models☆57Updated last month
- ☆85Updated last year
- ICML 2024 - Official Repository for EXO: Towards Efficient Exact Optimization of Language Model Alignment☆57Updated last year
- A Large-Scale, High-Quality Math Dataset for Reinforcement Learning in Language Models☆57Updated 3 months ago
- ☆114Updated 5 months ago
- Code for "Reasoning to Learn from Latent Thoughts"☆104Updated 2 months ago
- [ICLR 2025] Code for the paper "Implicit Search via Discrete Diffusion: A Study on Chess"☆27Updated 3 months ago
- [ICLR 2025] "Training LMs on Synthetic Edit Sequences Improves Code Synthesis" (Piterbarg, Pinto, Fergus)☆19Updated 4 months ago
- Repository for Skill Set Optimization☆13Updated 10 months ago
- ☆95Updated 11 months ago