Disentangling Optimization Scale from Preference Scale in DPO

arXiv:2608.27032 · cs.LG · Submitted 2026-08-27 · Read on arXiv

cs.LG

Submitted: 2026-08-27

Updated: 2026-08-27

Code: https://github.com/ivankru/bayesian_dpo

Terminology

• Adam — 19× in this paper · explained in 2 episodes
• Alignment — 14× in this paper · explained in 1 episode
• International Conference on Machine Learning — 9× in this paper
• Benchmark — 8× in this paper · explained in 2 episodes
• Direct Preference Optimization — 8× in this paper
• Entanglement — 5× in this paper · explained in 4 episodes
• Regularization — 5× in this paper · explained in 3 episodes
• Accuracy — 4× in this paper · explained in 1 episode
• Consistency — 4× in this paper · explained in 3 episodes
• Fine-Tuning — 4× in this paper · explained in 5 episodes
• LLM judge — 3× in this paper · explained in 2 episodes
• Direct Preference Optimization (DPO) — 2× in this paper · explained in 9 episodes
• Robustness — 2× in this paper · explained in 5 episodes
• Failure Modes — 1× in this paper · explained in 3 episodes
• In International Conference on Learning Representations — 1× in this paper
• Instruction Tuning — 1× in this paper · explained in 2 episodes
• Optimal Transport — 1× in this paper · explained in 5 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Reward Function — 1× in this paper · explained in 2 episodes
• Stochastic Dynamics — 1× in this paper · explained in 2 episodes
• Transformer — 1× in this paper · explained in 2 episodes

Sources

• A General Theoretical Paradigm to Understand Learning from Human Preferences
• Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
• Bootstrapping Language Models with DPO Implicit Rewards
• Length-Controlled AlpacaEval: A Simple Way to Debias Automatic Evaluators
• KTO: Model Alignment as Prospect Theoretic Optimization
• Geometric-Averaged Preference Optimization for Soft Preference Labels
• Explicit Preference Optimization: No Need for an Implicit Reward Model
• ROPO: Robust Preference Optimization for Large Language Models
• Gradient Imbalance in Direct Preference Optimization
• Distributional Preference Alignment of LLMs via Optimal Transport
• SimPO: Simple Preference Optimization with a Reference-Free Reward
• Entropy Controllable Direct Preference Optimization
• Direct Preference Optimization: Your Language Model is Secretly a Reward Model
• Provably Robust DPO: Aligning Language Models with Noisy Feedback
• Towards Data-Centric RLHF: Simple Metrics for Preference Dataset Comparison
• Efficiently Training Low-Curvature Neural Networks
• Rethinking Bradley-Terry Models in Preference-Based Reward Modeling: Foundations, Theory, and Alternatives
• Direct Preference Optimization with Rating Information: Practical Algorithms and Provable Gains
• Secrets of RLHF in Large Language Models Part II: Reward Modeling
• HelpSteer3-Preference: Open Human-Annotated Preference Data across Diverse Tasks and Languages

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