Dual Active Learning for Reinforcement Learning from Human Feedback
stat.ML, cs.LG
Submitted: 2024-10-03
Updated: 2026-11-06
Terminology
Sources
- A reinforced learning approach to optimal design under model uncertainty
- Comparing Rationality Between Large Language Models and Humans: Insights and Open Questions
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- Enhancing Causal Reasoning in Large Language Models: A Causal Attribution Model for Precision Fine-Tuning
- STEEL: Singularity-aware Reinforcement Learning
- Covariate Assisted Entity Ranking with Sparse Intrinsic Scores
- Active teacher selection for reward learning
- ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools
- Uncertainty in Language Models: Assessment through Rank-Calibration
- Reinforcement Learning from Human Feedback with Active Queries
- Policy learning "without" overlap: Pessimism and generalized empirical Bernstein's inequality
- Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems
- Robust Detection of Watermarks for Large Language Models Under Human Edits
- Sample Efficient Preference Alignment in LLMs via Active Exploration
- Open-LLM-Leaderboard: From Multi-choice to Open-style Questions for LLMs Evaluation, Benchmark, and Arena
- Synthetic Oversampling: Theory and A Practical Approach Using LLMs to Address Data Imbalance
- Proximal Policy Optimization Algorithms
- Gemma: Open Models Based on Gemini Research and Technology
- On the Algorithmic Bias of Aligning Large Language Models with RLHF: Preference Collapse and Matching Regularization
- DeepSpeed-Chat: Easy, Fast and Affordable RLHF Training of ChatGPT-like Models at All Scales
Related papers
- Uniform-in-Time Weak Propagation-of-Chaos in Shallow Neural Networks
- Diffusion Models Observe Only Gradients: A Geometric Perspective on Score Matching Errors
- Quantifying Error Propagation and Model Collapse in Diffusion Models
- Behavior of prediction performance metrics with rare events
- Statistical Inference and Learning for Shapley Additive Explanations (SHAP)
- Control Variate Design for Score-Function Gradient Estimators in Importance-Weighted Variational Inference