Uncovering and Mitigating Aggregation-Induced Reward Hacking in Multi-Reward Reinforcement Learning
cs.CL
Submitted: 2026-08-31
Updated: 2026-08-31
Code: https://github.com/yyhappier/AMRP
Terminology
Sources
- Gradient Regularization Mitigates Reward Hacking in Reinforcement Learning from Human Feedback and Verifiable Rewards
- Reward Model Ensembles Help Mitigate Overoptimization
- Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators
- Reward Shaping to Mitigate Reward Hacking in RLHF
- The Llama 3 Herd of Models
- From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline
- WARM: On the Benefits of Weight Averaged Reward Models
- Proximal Policy Optimization Algorithms
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Qwen3 Technical Report
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering