Towards Reliable, Uncertainty-Aware Alignment

• URL: http://arxiv.org/abs/2507.15906v1
• Date: Mon, 21 Jul 2025 12:51:29 GMT
• Title: Towards Reliable, Uncertainty-Aware Alignment
• Authors: Debangshu Banerjee, Kintan Saha, Aditya Gopalan,
• Abstract summary: We study the variability of reward model training on open-source benchmarks.<n>We propose a variance-aware policy optimization framework for preference-based alignment.
• Score: 12.63619480522393
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Alignment of large language models (LLMs) typically involves training a reward model on preference data, followed by policy optimization with respect to the reward model. However, optimizing policies with respect to a single reward model estimate can render it vulnerable to inaccuracies in the reward model. We empirically study the variability of reward model training on open-source benchmarks. We observe that independently trained reward models on the same preference dataset can exhibit substantial disagreement, highlighting the instability of current alignment strategies. Employing a theoretical model, we demonstrate that variability in reward model estimation can cause overfitting, leading to the risk of performance degradation. To mitigate this risk, we propose a variance-aware policy optimization framework for preference-based alignment. The key ingredient of the framework is a new policy regularizer that incorporates reward model variance estimates. We show that variance-aware policy optimization provably reduces the risk of outputting a worse policy than the default. Experiments across diverse LLM and reward model configurations confirm that our approach yields more stable and robust alignment than the standard (variance-unaware) pipeline.

Related papers

• Risk-aware Direct Preference Optimization under Nested Risk Measure [23.336246526648374]
  Risk-aware Direct Preference Optimization (Ra-DPO) is a novel approach that incorporates risk-awareness by employing a class of nested risk measures.<n> Experimental results across three open-source datasets demonstrate the proposed method's superior performance in balancing alignment performance and model drift.
  arXiv  Detail & Related papers  (2025-05-26T08:01:37Z)
• Efficient Safety Alignment of Large Language Models via Preference Re-ranking and Representation-based Reward Modeling [84.00480999255628]
  Reinforcement Learning algorithms for safety alignment of Large Language Models (LLMs) encounter the challenge of distribution shift.<n>Current approaches typically address this issue through online sampling from the target policy.<n>We propose a new framework that leverages the model's intrinsic safety judgment capability to extract reward signals.
  arXiv  Detail & Related papers  (2025-03-13T06:40:34Z)
• Towards Reliable Alignment: Uncertainty-aware RLHF [14.20181662644689]
  We show that the fluctuation of reward models can be detrimental to the alignment problem. We show that such policies are more risk-averse in the sense that they are more cautious of uncertain rewards. We use this ensemble of reward models to align language model using our methodology and observe that our empirical findings match our theoretical predictions.
  arXiv  Detail & Related papers  (2024-10-31T08:26:51Z)
• Self-Evolutionary Large Language Models through Uncertainty-Enhanced Preference Optimization [9.618391485742968]
  Iterative preference optimization has recently become one of the de-facto training paradigms for large language models (LLMs) We present an uncertainty-enhanced textbfPreference textbfOptimization framework to make the LLM self-evolve with reliable feedback. Our framework substantially alleviates the noisy problem and improves the performance of iterative preference optimization.
  arXiv  Detail & Related papers  (2024-09-17T14:05:58Z)
• Robust Preference Optimization through Reward Model Distillation [68.65844394615702]
  Direct Preference Optimization (DPO) is a popular offline alignment method that trains a policy directly on preference data.<n>We analyze this phenomenon and use distillation to get a better proxy for the true preference distribution over generation pairs.<n>Our results show that distilling from such a family of reward models leads to improved robustness to distribution shift in preference annotations.
  arXiv  Detail & Related papers  (2024-05-29T17:39:48Z)
• Provably Mitigating Overoptimization in RLHF: Your SFT Loss is Implicitly an Adversarial Regularizer [52.09480867526656]
  We identify the source of misalignment as a form of distributional shift and uncertainty in learning human preferences.<n>To mitigate overoptimization, we first propose a theoretical algorithm that chooses the best policy for an adversarially chosen reward model.<n>Using the equivalence between reward models and the corresponding optimal policy, the algorithm features a simple objective that combines a preference optimization loss and a supervised learning loss.
  arXiv  Detail & Related papers  (2024-05-26T05:38:50Z)
• Optimal Baseline Corrections for Off-Policy Contextual Bandits [61.740094604552475]
  We aim to learn decision policies that optimize an unbiased offline estimate of an online reward metric. We propose a single framework built on their equivalence in learning scenarios. Our framework enables us to characterize the variance-optimal unbiased estimator and provide a closed-form solution for it.
  arXiv  Detail & Related papers  (2024-05-09T12:52:22Z)
• Model-based Offline Policy Optimization with Adversarial Network [0.36868085124383626]
  We propose a novel Model-based Offline policy optimization framework with Adversarial Network (MOAN) Key idea is to use adversarial learning to build a transition model with better generalization. Our approach outperforms existing state-of-the-art baselines on widely studied offline RL benchmarks.
  arXiv  Detail & Related papers  (2023-09-05T11:49:33Z)
• Fine-Tuning Language Models with Advantage-Induced Policy Alignment [80.96507425217472]
  We propose a novel algorithm for aligning large language models to human preferences. We show that it consistently outperforms PPO in language tasks by a large margin. We also provide a theoretical justification supporting the design of our loss function.
  arXiv  Detail & Related papers  (2023-06-04T01:59:40Z)
• When Demonstrations Meet Generative World Models: A Maximum Likelihood Framework for Offline Inverse Reinforcement Learning [62.00672284480755]
  This paper aims to recover the structure of rewards and environment dynamics that underlie observed actions in a fixed, finite set of demonstrations from an expert agent. Accurate models of expertise in executing a task has applications in safety-sensitive applications such as clinical decision making and autonomous driving.
  arXiv  Detail & Related papers  (2023-02-15T04:14:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.