From Constitutions to Control: Interpretable Rewards for Aligning Language Models
cs.LG
Submitted: 2026-09-27
Updated: 2026-09-27
Code: https://github.com/jgaeb/rubric-rewards
Terminology
Sources
- gpt-oss-120b & gpt-oss-20b Model Card
- Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
- QA-LIGN: Aligning LLMs through Constitutionally Decomposed QA
- Quantile Regression for Distributional Reward Models in RLHF
- Gemma 3 Technical Report
- LoRA: Low-Rank Adaptation of Large Language Models
- Reinforcement Learning with Rubric Anchors
- Mitigating Label Bias with Interpretable Rubric Embeddings
- Interpretable Reward Modeling with Active Concept Bottlenecks
- Tulu 3: Pushing Frontiers in Open Language Model Post-Training
- Reward Bias Substitution: Single-Axis Bias Mitigations Redirect Optimization Pressure
- Reward Hacking in Rubric-Based Reinforcement Learning
- How RLHF Amplifies Sycophancy
- SteerRM: Debiasing Reward Models via Sparse Autoencoders
- Beyond Reward Hacking: Causal Rewards for Large Language Model Alignment
- Auto-Rubric: Learning From Implicit Weights to Explicit Rubrics for Reward Modeling
- Qwen3 Technical Report
- Rewards-in-Context: Multi-objective Alignment of Foundation Models with Dynamic Preference Adjustment
- Instruction-Following Evaluation for Large 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: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks