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.
☆32Apr 20, 2024Updated 2 years ago
Alternatives and similar repositories for understanding-rlhf
Users that are interested in understanding-rlhf are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Official Code Release for "Training a Generally Curious Agent"☆47May 18, 2025Updated last year
- Code for RATIONALYST: Pre-training Process-Supervision for Improving Reasoning https://arxiv.org/pdf/2410.01044☆35Oct 3, 2024Updated last year
- ☆25Apr 3, 2025Updated last year
- Multi-Agent Verification: Scaling Test-Time Compute with Multiple Verifiers☆29Mar 1, 2025Updated last year
- Official repo for Offline RL for Online RL☆18Oct 14, 2023Updated 2 years ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- ☆10Nov 17, 2022Updated 3 years ago
- ☆35Apr 12, 2024Updated 2 years ago
- Code for Diagnosing Bottlenecks in Deep Q-learning. Contains implementations of tabular environments plus solvers.☆17May 14, 2019Updated 7 years ago
- ☆15Oct 4, 2024Updated last year
- (NeurIPS '22) LISA: Learning Interpretable Skill Abstractions - A framework for unsupervised skill learning using Imitation☆29Feb 22, 2023Updated 3 years ago
- Optim4RL is a Jax framework of learning to optimize for reinforcement learning.☆28Nov 27, 2024Updated last year
- TACL 2025: Investigating Adversarial Trigger Transfer in Large Language Models☆19Aug 17, 2025Updated 9 months ago
- Anchored Preference Optimization and Contrastive Revisions: Addressing Underspecification in Alignment☆62Aug 30, 2024Updated last year
- ☆12Jul 4, 2024Updated last year
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- Training tiny models to prove hard theorems☆77Mar 5, 2026Updated 2 months ago
- ☆24Feb 16, 2022Updated 4 years ago
- ☆16Feb 22, 2025Updated last year
- Drop-in environment replacements that make your RL algorithm train faster.☆22Jun 19, 2024Updated last year
- ☆13Nov 15, 2022Updated 3 years ago
- This is code for most of the experiments in the paper Understanding the Effects of RLHF on LLM Generalisation and Diversity☆49Jan 19, 2024Updated 2 years ago
- Code for EMNLP 2021 paper "Measuring Association Between Labels and Free-Text Rationales"☆12Sep 12, 2023Updated 2 years ago
- RFTT: Reasoning with Reinforced Functional Token Tuning☆29Feb 12, 2026Updated 3 months ago
- ☆21Apr 3, 2026Updated last month
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- ☆36Jul 9, 2025Updated 10 months ago
- ☆99Jun 27, 2024Updated last year
- The official implementation of Self-Exploring Language Models (SELM)☆63Jun 4, 2024Updated last year
- Code to accompany the paper "The Information Geometry of Unsupervised Reinforcement Learning"☆20Oct 6, 2021Updated 4 years ago
- Self-Supervised Alignment with Mutual Information☆20May 24, 2024Updated 2 years ago
- Emergent Hierarchical Reasoning in LLMs/VLMs through Reinforcement Learning [ICLR26]☆64Apr 11, 2026Updated last month
- ☆14Jun 24, 2024Updated last year
- ☆29Feb 17, 2024Updated 2 years ago
- [ICLR 2025] Unintentional Unalignment: Likelihood Displacement in Direct Preference Optimization☆32Jan 7, 2026Updated 4 months ago
- Wordpress hosting with auto-scaling - Free Trial Offer • AdFully Managed hosting for WordPress and WooCommerce businesses that need reliable, auto-scalable performance. Cloudways SafeUpdates now available.
- Fork of Flame repo for training of some new stuff in development☆19Apr 24, 2026Updated last month
- ☆12Dec 22, 2021Updated 4 years ago
- Official implementation of the DECKARD Agent from the paper "Do Embodied Agents Dream of Pixelated Sheep?"☆94May 23, 2023Updated 3 years ago
- [COLM 2025] Official code for "When To Solve, When To Verify: Compute-Optimal Problem Solving and Generative Verification for LLM Reasoni…☆15Oct 31, 2025Updated 6 months ago
- Experiment for Understanding the Effects of Dataset Characteristics on Offline Reinforcement Learning☆26Jan 16, 2023Updated 3 years ago
- This repository contains the implementation of the PTR algorithm described in the paper: Pre-Training for Robots: Leveraging Diverse Mult…☆32Oct 26, 2022Updated 3 years ago
- [ICLR 2026] PSFT is a trust-region–inspired fine-tuning objective that views SFT as a policy gradient method with constant advantages, co…☆38Sep 9, 2025Updated 8 months ago