AERO: Adaptive Ensemble-Disagreement Routing for Oracle Feedback in Sample-Efficient Online RLHF
cs.LG
Submitted: 2026-11-07
Updated: 2026-11-07
Code: https://github.com/Shrey-Singhal/AERO
Terminology
Sources
- Efficient Exploration at Scale
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- The Llama 3 Herd of Models
- Direct Language Model Alignment from Online AI Feedback
- Bayesian Active Learning for Classification and Preference Learning
- Skywork-Reward: Bag of Tricks for Reward Modeling in LLMs
- Sample-Efficient Alignment for LLMs
- Proximal Policy Optimization Algorithms
- Understanding the performance gap between online and offline alignment algorithms
- Some things are more CRINGE than others: Iterative Preference Optimization with the Pairwise Cringe Loss
- Qwen2.5 Technical Report
- Self-Exploring Language Models: Active Preference Elicitation for Online Alignment
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis