Greedy Sampling Is Provably Efficient for RLHF
"人类反馈" 相关文章
Improving Video Generation with Human Feedback
RobotArena $\infty$: Scalable Robot Benchmarking via Real-to-Sim Translation
Rectifying Shortcut Behaviors in Preference-based Reward Learning
Offline and Online KL-Regularized RLHF under Differential Privacy
Provably Mitigating Corruption, Overoptimization, and Verbosity Simultaneously in Offline and Online RLHF/DPO Alignment
General Exploratory Bonus for Optimistic Exploration in RLHF
Rethinking KL Regularization in RLHF: From Value Estimation to Gradient Optimization
Feedback Forensics: A Toolkit to Measure AI Personality
Avoiding $\mathbf{exp(R_{max})}$ scaling in RLHF through Preference-based Exploration
RLBFF: Binary Flexible Feedback to bridge between Human Feedback & Verifiable Rewards
AI Alignment Is Trivial
Opal: An Operator Algebra View of RLHF
研究人员提出AI对齐新方法，通过“交互式分解”改善人类提供反馈过程
Fine-tuning GPT-2 from human preferences
从原理到实战：RLHF（人类反馈强化学习）完整流程
Ambiguity Resolution with Human Feedback for Code Writing Tasks
Inclusion Arena: An Open Platform for Evaluating Large Foundation Models with Real-World Apps
Forest vs Tree: The $(N, K)$ Trade-off in Reproducible ML Evaluation
Augmented Reinforcement Learning Framework For Enhancing Decision-Making In Machine Learning Models Using External Agents
