Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics
cs.LG, cs.AI
Submitted: 2026-09-30
Updated: 2026-09-30
Code: https://github.com/EQ-bench/creative-writing-bench
Terminology
Sources
- HealthBench: Evaluating Large Language Models Towards Improved Human Health
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- Constitutional AI: Harmlessness from AI Feedback
- Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators
- Helping or Herding? Reward Model Ensembles Mitigate but do not Eliminate Reward Hacking
- HealthBench Professional: Evaluating Large Language Models on Real Clinician Chats
- RISE-RL: Rubric-Informed Selective Exploration for Open-Ended Reinforcement Learning
- Reinforcement Learning with Rubric Anchors
- Open Rubric System: Scaling Reinforcement Learning with Pairwise Adaptive Rubric
- Prometheus: Inducing Fine-grained Evaluation Capability in Language Models
- RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback
- From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline
- OpenRubrics: Towards Scalable Synthetic Rubric Generation for Reward Modeling and LLM Alignment
- Mitigating False Credit Propagation: Probabilistic Graphical Reward Aggregation for Rubric-Based Reinforcement Learning
- Reward Hacking in Rubric-Based Reinforcement Learning
- Training language models to follow instructions with human feedback
- Scaling Laws for Reward Model Overoptimization in Direct Alignment Algorithms
- GPQA: A Graduate-Level Google-Proof Q&A Benchmark
- Rubric-Guided Self-Distillation: Post-Training Without Rubric Verifiers
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks