Targeting Misalignment: A Conflict-Aware Framework for Reward-Model-based LLM Alignment

• URL: http://arxiv.org/abs/2512.09212v1
• Date: Wed, 10 Dec 2025 00:52:21 GMT
• Title: Targeting Misalignment: A Conflict-Aware Framework for Reward-Model-based LLM Alignment
• Authors: Zixuan Liu, Siavash H. Khajavi, Guangkai Jiang, Xinru Liu,
• Abstract summary: Reward-model-based fine-tuning is a central paradigm in aligning Large Language Models with human preferences.<n>This paper investigates a novel framework to identify and mitigate such misalignment by treating the fine-tuning process as a form of knowledge integration.
• Score: 5.900494456937422
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Reward-model-based fine-tuning is a central paradigm in aligning Large Language Models with human preferences. However, such approaches critically rely on the assumption that proxy reward models accurately reflect intended supervision, a condition often violated due to annotation noise, bias, or limited coverage. This misalignment can lead to undesirable behaviors, where models optimize for flawed signals rather than true human values. In this paper, we investigate a novel framework to identify and mitigate such misalignment by treating the fine-tuning process as a form of knowledge integration. We focus on detecting instances of proxy-policy conflicts, cases where the base model strongly disagrees with the proxy. We argue that such conflicts often signify areas of shared ignorance, where neither the policy nor the reward model possesses sufficient knowledge, making them especially susceptible to misalignment. To this end, we propose two complementary metrics for identifying these conflicts: a localized Proxy-Policy Alignment Conflict Score (PACS) and a global Kendall-Tau Distance measure. Building on this insight, we design an algorithm named Selective Human-in-the-loop Feedback via Conflict-Aware Sampling (SHF-CAS) that targets high-conflict QA pairs for additional feedback, refining both the reward model and policy efficiently. Experiments on two alignment tasks demonstrate that our approach enhances general alignment performance, even when trained with a biased proxy reward. Our work provides a new lens for interpreting alignment failures and offers a principled pathway for targeted refinement in LLM training.

Related papers

• OmniVL-Guard: Towards Unified Vision-Language Forgery Detection and Grounding via Balanced RL [63.388513841293616]
  Existing forgery detection methods fail to handle the interleaved text, images, and videos prevalent in real-world misinformation.<n>To bridge this gap, this paper targets to develop a unified framework for omnibus vision-language forgery detection and grounding.<n>We propose textbf OmniVL-Guard, a balanced reinforcement learning framework for omnibus vision-language forgery detection and grounding.
  arXiv  Detail & Related papers  (2026-02-11T09:41:36Z)
• Search-R2: Enhancing Search-Integrated Reasoning via Actor-Refiner Collaboration [49.9937230730202]
  We propose Search-R2, a novel Actor-Refiner collaboration framework that enhances reasoning through targeted intervention.<n>Our approach decomposes the generation process into an Actor, which produces initial reasoning trajectories.<n>We show that Search-R2 consistently outperforms strong RAG and RL-based baselines across model scales.
  arXiv  Detail & Related papers  (2026-02-03T15:32:09Z)
• Reward-free Alignment for Conflicting Objectives [12.275610380458119]
  We propose a Reward-free Alignment framework for Conflicted Objectives (RACO)<n>RACO directly leverages pairwise preference data and resolves gradient conflicts via a novel clipped variant of conflict-averse gradient descent.<n>We provide convergence guarantees to Pareto-critical points that respect user-specified objective weights, and further show that clipping can strictly improve convergence rate in the two-objective setting.
  arXiv  Detail & Related papers  (2026-02-02T18:59:52Z)
• How RLHF Amplifies Sycophancy [23.213056717401418]
  Large language models often exhibit increased sycophantic behavior after preference-based post-training.<n>We identify an explicit amplification mechanism that causally links optimization against a learned reward to bias in the human preference data used for alignment.<n>We propose a training-time intervention designed to neutralize the amplification mechanism itself.
  arXiv  Detail & Related papers  (2026-02-01T03:46:14Z)
• Are LLMs Vulnerable to Preference-Undermining Attacks (PUA)? A Factorial Analysis Methodology for Diagnosing the Trade-off between Preference Alignment and Real-World Validity [45.92643973404507]
  We investigate whether aligned models are vulnerable to Preference-Undermining Attacks (PUA), a class of manipulative prompting strategies.<n>Surprisingly, more advanced models are sometimes more susceptible to manipulative prompts.
  arXiv  Detail & Related papers  (2026-01-10T15:16:23Z)
• Maestro: Learning to Collaborate via Conditional Listwise Policy Optimization for Multi-Agent LLMs [23.590034731179824]
  We present Through Role Orchestration (Maestro), a principled paradigm for collaboration that structurally decouples cognitive modes.<n>Maestro uses a collective of parallel Execution Agents for diverse exploration and a specialized Central Agent for convergent, evaluative synthesis.<n>Experiments on mathematical reasoning and general problem-solving benchmarks demonstrate that Maestro, coupled with CLPO, consistently outperforms existing state-of-the-art multi-agent approaches.
  arXiv  Detail & Related papers  (2025-11-08T21:01:27Z)
• Taming the Judge: Deconflicting AI Feedback for Stable Reinforcement Learning [46.661195064495]
  This work introduces an end to end framework to detect and resolve inconsistencies within the reinforcement learning training loop.<n>Our framework features two core contributions: the Conflict Detection Rate (CDR), a novel metric to quantify judgment conflicts, and Deconflicted Graph Rewards (DGR), a signal-purification framework.<n>Experiments confirm that our framework significantly improves training stability and model performance over strong baselines.
  arXiv  Detail & Related papers  (2025-10-17T10:34:59Z)
• Beyond Reward Hacking: Causal Rewards for Large Language Model Alignment [30.605500809158986]
  We propose a novel causal reward modeling approach that integrates causality to mitigate spurious correlations.<n>Our approach mitigates various types of spurious correlations effectively, resulting in more reliable and fair alignment of LLMs with human preferences.
  arXiv  Detail & Related papers  (2025-01-16T16:00:37Z)
• Mitigating Distribution Shift in Model-based Offline RL via Shifts-aware Reward Learning [36.01269673940484]
  This paper offers a comprehensive analysis that disentangles the problem into two fundamental components: model bias and policy shift.<n>Our theoretical and empirical investigations reveal how these factors distort value estimation and policy optimization.<n>We derive a novel shifts-aware reward through a unified probabilistic inference framework, which modifies the vanilla reward to refine value learning and facilitate policy training.
  arXiv  Detail & Related papers  (2024-08-23T04:25:09Z)
• Learning Reward and Policy Jointly from Demonstration and Preference Improves Alignment [58.049113055986375]
  We develop a single stage approach named Alignment with Integrated Human Feedback (AIHF) to train reward models and the policy.<n>The proposed approach admits a suite of efficient algorithms, which can easily reduce to, and leverage, popular alignment algorithms.<n>We demonstrate the efficiency of the proposed solutions with extensive experiments involving alignment problems in LLMs and robotic control problems in MuJoCo.
  arXiv  Detail & Related papers  (2024-06-11T01:20:53Z)
• Learn from the Past: A Proxy Guided Adversarial Defense Framework with Self Distillation Regularization [53.04697800214848]
  Adversarial Training (AT) is pivotal in fortifying the robustness of deep learning models. AT methods, relying on direct iterative updates for target model's defense, frequently encounter obstacles such as unstable training and catastrophic overfitting. We present a general proxy guided defense framework, LAST' (bf Learn from the Pbf ast)
  arXiv  Detail & Related papers  (2023-10-19T13:13:41Z)
• PARL: A Unified Framework for Policy Alignment in Reinforcement Learning from Human Feedback [106.63518036538163]
  We present a novel unified bilevel optimization-based framework, textsfPARL, formulated to address the recently highlighted critical issue of policy alignment in reinforcement learning. Our framework addressed these concerns by explicitly parameterizing the distribution of the upper alignment objective (reward design) by the lower optimal variable. Our empirical results substantiate that the proposed textsfPARL can address the alignment concerns in RL by showing significant improvements.
  arXiv  Detail & Related papers  (2023-08-03T18:03:44Z)
• Provable Offline Preference-Based Reinforcement Learning [95.00042541409901]
  We investigate the problem of offline Preference-based Reinforcement Learning (PbRL) with human feedback. We consider the general reward setting where the reward can be defined over the whole trajectory. We introduce a new single-policy concentrability coefficient, which can be upper bounded by the per-trajectory concentrability.
  arXiv  Detail & Related papers  (2023-05-24T07:11:26Z)
• Effective Targeted Attacks for Adversarial Self-Supervised Learning [58.14233572578723]
  unsupervised adversarial training (AT) has been highlighted as a means of achieving robustness in models without any label information. We propose a novel positive mining for targeted adversarial attack to generate effective adversaries for adversarial SSL frameworks. Our method demonstrates significant enhancements in robustness when applied to non-contrastive SSL frameworks, and less but consistent robustness improvements with contrastive SSL frameworks.
  arXiv  Detail & Related papers  (2022-10-19T11:43:39Z)
• Resisting Adversarial Attacks in Deep Neural Networks using Diverse Decision Boundaries [12.312877365123267]
  Deep learning systems are vulnerable to crafted adversarial examples, which may be imperceptible to the human eye, but can lead the model to misclassify. We develop a new ensemble-based solution that constructs defender models with diverse decision boundaries with respect to the original model. We present extensive experimentations using standard image classification datasets, namely MNIST, CIFAR-10 and CIFAR-100 against state-of-the-art adversarial attacks.
  arXiv  Detail & Related papers  (2022-08-18T08:19:26Z)
• Formalizing the Problem of Side Effect Regularization [81.97441214404247]
  We propose a formal criterion for side effect regularization via the assistance game framework. In these games, the agent solves a partially observable Markov decision process. We show that this POMDP is solved by trading off the proxy reward with the agent's ability to achieve a range of future tasks.
  arXiv  Detail & Related papers  (2022-06-23T16:36:13Z)

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.