Preference Alignment Learning

Follow this topic Ask about this topic

Momentum

52 papers in the last four weeks, up 333% on the four weeks before. 0.5% of all new papers.

Jul 13Week of Sep 28

Latest papers 271

May 19, 2026cs.GT

PrefBench: Evaluating Zero-Shot LLM Agents in Hidden-Preference Personalized Pricing Negotiations

Personalized pricing negotiations are a challenging testbed for LLM agents because successful interaction does not guarantee profitable decision making. A seller may produce valid actions and close many deals while still pricing poorly when buyer willingness to pay and bargaining traits remain hidden. This paper presents PrefBench, a simulator-based benchmark for hidden-preference personalized pricing negotiations. Each episode pairs a simulated buyer with a fixed vehicle-customization bundle; the seller observes public persona descriptors, bundle information, and negotiation history, while latent buyer variables govern valuation, patience, counter-offer behavior, and walkaway decisions. PrefBench evaluates this setting through an LLM-facing state-summary protocol that constrains agents to return strict JSON actions under a fixed hidden-information boundary. We evaluate zero-shot LLM sellers against heuristic references over 7,500 episodes. The tested LLMs follow the protocol reliably and achieve deal rates above 0.99, but their seller-profit outcomes remain weak: the best LLM average profit is only slightly above the random baseline and far below a simple concession heuristic under the same episode stream. These results show that structured action compliance and agreement-seeking behavior can coexist with weak profit-sensitive bargaining. PrefBench provides a controlled benchmark for evaluating pricing-agent behavior under hidden buyer preferences.

Yingjie Lei

University of Aberdeen

Negotiation Agentic Commerce Large Language Model Agents Preference Alignment Learning Refusals

May 18, 2026cs.LG

General Preference Reinforcement Learning

Post-training has split large language model (LLM) alignment into two largely disconnected tracks. Online reinforcement learning (RL) with verifiable rewards drives emergent reasoning on math and code but depends on a programmatic verifier that cannot reach open-ended tasks, while preference optimization handles open-ended generation yet forgoes the continuous exploration that powers online RL. Closing this gap requires a verifier for open-ended quality, but a scalar reward model is the wrong shape for the job. Quality is multi-dimensional, and any scalar score is an incomplete proxy that lets online RL collapse onto whichever axis the score is most sensitive to. We turn instead to the General Preference Model (GPM), which embeds responses into

k

skew-symmetric subspaces and represents preference as a structured, intransitivity-aware comparison. Building on this, we propose General Preference Reinforcement Learning (GPRL), which carries the

k

-way structure through to the policy update. GPRL computes per-dimension group-relative advantages, normalizes each on its own scale so no axis can dominate, and aggregates them with context-dependent eigenvalues. The same structure powers a closed-loop drift monitor that detects single-axis exploitation and corrects it on the fly by reweighting dimensions and tightening the trust region. Starting from

\texttt{Llama-3-8B-Instruct}

, GPRL reaches a length-controlled win rate of

56.51\%

on AlpacaEval~2.0 while also outperforming SimPO and SPPO on Arena-Hard, MT-Bench, and WildBench by resisting reward hacking across extended training runs.

Muhammad Umer, Muhammad Ahmed Mohsin, Ahsan Bilal +5

Stanford University · The University of Oklahoma

Large Language Model Reinforcement Learning Preference Alignment Learning

May 18, 2026cs.LG

GAMMA: Global Bit Allocation for Mixed-Precision Models under Arbitrary Budgets

Mixed-precision quantization improves the budget--accuracy trade-off for large language models (LLMs) by allocating more bits to sensitive modules. However, automating this allocation at LLM scale faces a unique combination of constraints: learnable approaches require quantization-aware training, which is infeasible for billion-parameter models; training-free alternatives rely on static proxy metrics that miss cross-module interactions and must be recomputed per target budget; and search-based methods are expensive without guaranteeing exact budget compliance. We propose GAMMA, a quantizer-agnostic framework that learns module-wise precision preferences entirely within a post-training pipeline. GAMMA optimizes a teacher-forced hidden-state reconstruction objective under an augmented Lagrangian constraint, and projects the learned preferences into exact budget-feasible discrete assignments via integer programming. A key property is score reuse: because the learned preferences encode a stable sensitivity ranking rather than budget-specific weights, a single training run serves arbitrary deployment targets by re-solving only the integer program, reducing per-budget adaptation from hours to a few minutes. Across Llama and Qwen models (8B--32B), GAMMA outperforms both fixed-precision baselines (up to +12.99 Avg.) and search-based mixed-precision methods (up to +7.00 Avg.), and can match fixed 3-bit quality at 2.5-bit average precision, enabling deployment at substantially smaller memory footprints.

Zhangyang Yao, Haiyan Zhao, Haoyu Wang +3

Beihang University · Tsinghua University · ByteDance Inc

Mixed-Precision Quantization Token Budget Allocation Preference Alignment Learning

May 17, 2026cs.LG

ClaHF: A Human Feedback-inspired Reinforcement Learning Framework for Improving Classification Tasks

Text classification models are typically trained via supervised fine-tuning (SFT). However, SFT essentially performs behavior cloning from instance-wise labels and thus fails to adequately capture relative preference relations among samples, which limits the model's ability to shape decision boundaries and calibrate predictive confidence. In this paper, we propose ClaHF, a human feedback-inspired reinforcement learning (RL) framework for text classification that integrates preference modeling and RL optimization into the classification pipeline without requiring additional human annotations. Unlike prior work that relies solely on instance-wise supervision, ClaHF constructs multiple candidate predictions together with their relative ranking relations, and jointly models the Top-1 preference and the ordering among non-optimal candidates within a reward model (RM). This design converts conventional label supervision into preference signals that are directly applicable to policy optimization. We conduct systematic evaluations on eight classification tasks spanning three categories of scenarios. Results demonstrate that ClaHF consistently improves both classification performance and confidence calibration across diverse language models (LMs). The data and code are available at https://anonymous.4open.science/r/ClaHF.

Tianxiang Xu, Xiaoyan Zhu, Xin Lai +1

School of Computer Science and Technology, Xi’an Jiaotong University

Reinforcement Learning From Human Feedback Large Language Model Classification Preference Alignment Learning

May 17, 2026cs.CL

Transitivity Meets Cyclicity: Explicit Preference Decomposition for Dynamic Large Language Model Alignment

Standard RLHF relies on transitive scalar rewards, failing to capture the cyclic nature of human preferences. While some approaches like the General Preference Model (GPM) address this, we identify a theoretical limitation: their implicit formulation entangles hierarchy with cyclicity, failing to guarantee dominant solutions. To address this, we propose the Hybrid Reward-Cyclic (HRC) model, which utilizes game-theoretic decomposition to explicitly disentangle preferences into orthogonal transitive (scalar) and cyclic (vector) components. Complementing this, we introduce Dynamic Self-Play Preference Optimization (DSPPO), which treats alignment as a time-varying game to progressively guide the policy toward the Nash equilibrium. Synthetic data experiments further validate HRC's structural superiority in mixed transitive--cyclic settings, where HRC converges faster and achieves higher accuracy than GPM. Experiments on RewardBench 2 demonstrate that HRC consistently improves over both BT and GPM baselines (e.g., +1.23% on Gemma-2B-it). In particular, its superior performance in the Ties domain empirically validates the model's robustness in handling complex, non-strict preferences. Extensive downstream evaluations on AlpacaEval 2.0, Arena-Hard-v0.1, and MT-Bench confirm the efficacy of our framework. Notably, when using Gemma-2B-it as the base preference model, HRC+DSPPO achieves a peak length-controlled win-rate of 44.75% on AlpacaEval 2.0 and 46.8% on Arena-Hard-v0.1, significantly outperforming SPPO baselines trained with BT or GPM. Our code is publicly available at https://github.com/lab-klc/Hybrid-Reward-Cyclic.

Yucong Huang, Xiucheng Li, Kaiqi Zhao +1

Harbin Institute of Technology, Shenzhen, China.

Preference Alignment Large Language Model Alignment Reinforcement Learning From Human Feedback Reward Functions Preference Alignment Learning Self-Play

May 15, 2026cs.LG

Learning What Evaluators Value: A Reliable Approach to Modeling Evaluator Preferences

In many applications, human and LLM evaluators use assessments of relevant criteria to create an overall evaluation for an item or individual. For example, in admissions, committees assess candidates on attributes such as test scores, GPA, and research experience to evaluate their overall fit for the program. Another example arises in medical care where clinicians use patient reports of symptoms to consider preliminary diagnoses and assess risks. Each setting involves mapping multiple criteria to an overall evaluation -- a process that reflects the evaluator's underlying preferences. We focus on the fundamental question of learning these preferences. Many applications of this problem make specific modeling assumptions on evaluator preferences that may be substantially violated in the real world. We make the minimal assumption that the preference function is coordinate-wise non-decreasing, which is reasonable in a large number of evaluation settings. We theoretically characterize the severity of model mismatch for many common assumptions and show that it can lead to significant issues for learning evaluator preferences and other important downstream tasks. We then present an algorithm for learning evaluators' preferences that is robust to model mismatch. We prove theoretically that our algorithm can learn any preference function without sacrificing performance when the linearity assumption holds. Evaluations of our algorithm with synthetic simulations and real-world data confirm its ability to learn preferences robustly and illustrate key aspects of LLM and human preferences.

Madeline Celi Kitch, Nihar B. Shah

Carnegie Mellon University

Preference Alignment Learning Rater Mismatch Reliable

May 15, 2026cs.RO

OHP-RL: Online Human Preference as Guidance in Reinforcement Learning for Robot Manipulation

While reinforcement learning (RL) enables robots to acquire skills autonomously, its real-world deployment is severely limited by inefficient and unsafe exploration. Human-in-the-loop interventions offer a practical solution, yet existing methods typically exploit these interventions as auxiliary training signals, without fully capturing the richer information they provide about when and how autonomy should be guided. Human interventions often encode relative preferences over behavior under safety and task constraints, rather than prescribing exact actions to imitate. Motivated by this perspective, we propose Online Human Preference as Guidance in Reinforcement Learning (OHP-RL), a framework that leverages human interventions as preference information to guide policy learning. OHP-RL introduces a state-dependent preference gate that adaptively regulates when and to what extent human interventions should shape policy learning. This design enables the agent to benefit from intermittent and imperfect human feedback while preserving autonomous exploration and stable policy optimization. We evaluate OHP-RL on three challenging real-world contact-rich manipulation tasks on a Franka robot. Across all tasks, OHP-RL consistently achieves strong success rates, faster convergence, and substantially lower human intervention effort than prior approaches. Moreover, the learned policies exhibit more stable and human-aligned behavior throughout training.

Yunyang Mo, Jian Li, Qiwei Wu +2

The Hong Kong University of Science and Technology (Guangzhou)

Reinforcement Learning From Human Feedback Robotic Manipulation Frictive Policy Optimization Preference Alignment Learning Guidance

May 15, 2026cs.CL

Contexting as Recommendation: Evolutionary Collaborative Filtering for Context Engineering

Large Language Models (LLMs) are highly sensitive to their input contexts, motivating the development of automated context engineering. However, existing methods predominantly treat this as a global search problem, seeking a single context strategy that maximizes average performance across a dataset. This restrictive assumption overlooks the fact that different inputs often require distinct guidance, leaving substantial instance-level performance gains untapped. In this paper, we propose a paradigm shift by formulating context engineering as a recommendation problem. We introduce \textbf{Neural Collaborative Context Engineering (NCCE)}, a framework that transitions optimization from a static global search to dynamic, instance-wise routing. NCCE first bootstraps a diverse catalog of anchor contexts and then employs a novel \textbf{Context-CF Co-Evolution} mechanism. This stage establishes a synergistic feedback loop: a lightweight Neural Collaborative Filtering (NCF) model learns instance-context preferences to guide the generation of specialized context variants, while the newly evaluated contexts continuously refine the NCF model's understanding of latent preferences. At inference time, the trained NCF model acts as a context router, dynamically assigning the most suitable context strategy to each unseen instance. Theoretical Proofs and comprehensive experiments demonstrate that by matching individual inputs with their optimal contexts, NCCE significantly improves task accuracy, highlighting the critical importance of personalization in LLM context engineering.

Jiachen Zhu, Zhuoying Ou, Congmin Zheng +9

Shanghai Jiao Tong Univ. · Univ. College London · Carnegie Mellon Univ. +1

Collaborative Filtering Recommendation Preference Alignment Learning Engineering

May 14, 2026cs.LG

Active Learners as Efficient PRP Rerankers

Pairwise Ranking Prompting (PRP) elicits pairwise preference judgments from an LLM, which are then aggregated into a ranking, usually via classical sorting algorithms. However, judgments are noisy, order-sensitive, and sometimes intransitive, so sorting assumptions do not match the setting. Because sorting aims to recover a full permutation, truncating it to meet a call budget does not produce a dependable top-K. We thus reframe PRP reranking as active learning from noisy pairwise comparisons and show that active rankers are drop-in replacements that improve NDCG@10 per call in the call-constrained regime. Our noise-robust framework also introduces a randomized-direction oracle that uses a single LLM call per pair. This approach converts systematic position bias into zero-mean noise, enabling unbiased aggregate ranking without the cost of bidirectional calls.

Jeremías Figueiredo Paschmann, Juan Kaplan, Francisco Nattero +3

ELIAS Lab, Departamento de Ingeniería, Universidad de San Andrés

Cross-Encoder Reranking Ranking Active Learning Preference Alignment Learning Prompt Engineering

May 13, 2026cs.CL

Direct Preference Optimization for English-Mandarin Code-Switching Speech Recognition in Audio LLMs

Audio large language models (Audio LLMs) exhibit systematic failures in transcribing code-switching speech despite strong multilingual capabilities. Focusing on English-Mandarin, we identify three failure modes: language omission, translation-instead-of-transcription, and hallucination. We apply Direct Preference Optimization (DPO) to align models, constructing preference pairs in which chosen responses preserve mixed-language content while rejected responses mimic failure patterns. Training three Audio LLMs on 100K pairs (570 hours), we observe consistent behavioral shifts: models learn to preserve language composition rather than translating when prompted for transcription. This alignment yields MER reductions up to 89.6% (in-distribution) and 20.0% (out-of-distribution). Our findings suggest DPO can effectively elicit correct code-switching transcription behavior from multilingual Audio LLMs.

Trung Nguyen Quang, Cheng Yi Lewis Won, Minh Duc Pham +3

Institute for Infocomm Research (I2R), A⋆STAR, Singapore · Nanyang Technological University, Singapore

Code-Switching Large Audio Language Models Mandarin Preference Alignment Learning

May 13, 2026cs.CL

Probing Persona-Dependent Preferences in Language Models

Large language models (LLMs) can be said to have preferences: they reliably pick certain tasks and outputs over others, and preferences shaped by post-training and prompting appear to influence much of their behaviour. But models can also adopt different personas which have radically different preferences. How is this implemented internally? Does each persona use its own preference representations, or are some representations shared? We train linear probes on residual-stream activations of Gemma-3-27B and Qwen-3.5-122B to predict revealed pairwise task choices, and identify a genuine preference vector: it tracks the model's preferences as they shift across a range of prompts and situations, and on Gemma-3-27B steering along it causally controls pairwise choice. Some preference information transfers across the prompted personas we test: a probe trained on the helpful assistant predicts and steers the choices of qualitatively different personas, including an evil persona whose preferences anti-correlate with the Assistant's.

Oscar Gilg, Pierre Beckmann, Daniel Paleka +1

MATS · MATS, EPFL · ETH Zürich +1

Preference Alignment Learning

May 13, 2026cs.AI

Improving Code Translation with Syntax-Guided and Semantic-aware Preference Optimization

LLMs have shown immense potential for code translation, yet they often struggle to ensure both syntactic correctness and semantic consistency. While preference-based learning offers a promising alignment strategy, it is hindered by unreliable semantic rewards derived from sparse test cases or restrictive reference translations. We argue that a robust semantic reward for code translation must be derived directly from the source code. In this paper, we propose CTO to improve code translation with syntax-guided and semantic-aware preference optimization. Through contrastive learning, we train a cross-lingual semantic model to directly assess functional equivalence between source and translated code. By formulating code translation as a multi-objective optimization problem, this robust semantic signal is seamlessly unified with compiler-based syntactic feedback within the direct preference optimization framework. Extensive experiments on C++, Java, and Python translations demonstrate that CTO significantly outperforms existing baselines and alternative preference optimization strategies.

Yuhan Wu, Huan Zhang, Wei Cheng +3

State Key Laboratory for Novel Software Technology, Nanjing University, China · National Institute of Healthcare Data Science, Nanjing University, China

Code Generation Preference Alignment Learning

May 12, 2026cs.AI

Learning Transferable Latent User Preferences for Human-Aligned Decision Making

Large language models (LLMs) are increasingly used as reasoning modules in many applications. While they are efficient in certain tasks, LLMs often struggle to produce human-aligned solutions. Human-aligned decision making requires accounting for both explicitly stated goals and latent user preferences that shape how ambiguous situations should be resolved. Existing approaches to incorporating such preferences either rely on extensive and repeated user interactions or fail to generalize latent preferences across tasks and contexts, limiting their practical applicability. We consider a setting in which an LLM is used for high-level reasoning and is responsible for inferring latent user preferences from limited interactions, which guides downstream decision making. We introduce CLIPR (Conversational Learning for Inferring Preferences and Reasoning), a framework that learns actionable, transferable natural language rules that represent latent user preferences from minimal conversational input. These rules are iteratively refined through adaptive feedback and applied to both in-distribution and out-of-distribution ambiguous tasks across multiple environments. Evaluations on three datasets and a user study show that CLIPR consistently outperforms existing methods in improving alignment and reducing inference costs.

Alina Hyk, Sandhya Saisubramanian

Oregon State University

Large Language Model Decisions Preference Alignment Learning

May 12, 2026stat.ML

Variance-aware Reward Modeling with Anchor Guidance

Standard Bradley--Terry (BT) reward models are limited when human preferences are pluralistic. Although soft preference labels preserve disagreement information, BT can only express it by shrinking reward margins. Gaussian reward models provide an alternative by jointly predicting a reward mean and a reward variance, but suffer from a fundamental non-identifiability from pairwise preferences alone. We propose Anchor-guided Variance-aware Reward Modeling, a framework that resolves this non-identifiability by augmenting preference data with two coarse response-level anchor labels. Building on this, we prove that two anchors are sufficient for identification, develop a joint training objective and establish a non-asymptotic convergence rate for both the estimated reward mean and variance functions. Across simulation studies and four real-world diverging-preference datasets, our method consistently improves reward modeling performance and downstream RLHF, including PPO training and best-of-

N

selection.

Shuxing Fang, Ruijian Han, Liangyu Zhang +1

Department of Data Science and Artificial Intelligence, Hong Kong Polytechnic University · School of Statistics and Data Science, Shanghai University of Finance and Economics

Progress Reward Modeling Preference Alignment Learning Guidance

May 12, 2026cs.AI

Explaining and Breaking the Safety-Helpfulness Ceiling via Preference Dimensional Expansion

In the realm of multi-objective alignment for large language models, balancing disparate human preferences often manifests as a zero-sum conflict. Specifically, the intrinsic tension between competing goals dictates that aggressively optimizing for one metric (e.g., helpfulness) frequently incurs a substantial penalty on another (e.g., harmlessness). While prior work mainly focuses on data selection, parameter merging, or algorithmic balancing during training, these approaches merely force compromises between divergent preferences along a fixed Pareto frontier, failing to fundamentally resolve the inherent trade-off. In this work, we approach this problem from a novel perspective of multi-dimensional rewards. By scaling up the model's rollouts and analyzing the outputs across different reward dimensions, we arrive at a critical conclusion: the conflict among multiple objectives stems from the fact that the prompt itself inherently restricts the achievable multi-dimensional rewards. Based on this core observation, we propose MORA: Multi-Objective Reward Assimilation. Specifically, MORA isolates single-reward prompts through pre-sampling and expands their reward diversity by rewriting the original questions to incorporate multi-dimensional intents. Extensive experiments demonstrate that: (1) in sequential alignment, MORA achieves single-preference improvements ranging from 5% to 12.4%, with exceptional gains in harmlessness, after multiple-preference alignment across helpful, harmless, and truthful dimensions. (2) In simultaneous alignment, MORA achieves an average overall reward improvement of 4.6%. Our codes are available at https://github.com/Shiying-Huang/MORA-MPA.

ShiYing Huang, Liang Lin, Yuer Li +6

Huazhong University of Science and Technology · Nanyang Technological University · Tsinghua University +1

Preference Alignment Multi-Objective Reinforcement Learning Preference Alignment Learning Reward Functions Ceiling Explainable AI Methods

May 11, 2026cs.LG

A Theory of Time-Sensitive Language Generation: Sparse Hallucination Beats Mode Collapse

We study language generation in the limit under a global preference ordering on strings, as introduced by Kleinberg and Wei. As is done in previous work, we aim for breadth, but impose an additional requirement of timeliness: higher-ranked strings should be generated earlier. A string is then only credited if it is generated before a deadline, where its deadline is defined by a function that maps a string's rank in the target language to the time by which it must be produced. This is in keeping with a central consideration in machine learning, where inductive bias favors simpler'' or more plausible'' outputs, all else being equal. We show that timely generation is impossible in a strong sense for eventually consistent generators -- the protagonists of most prior related work. Under what is perhaps the mildest natural relaxation of consistency, a hallucination rate that vanishes over time, we show that we can circumvent our impossibility result. In particular, we can achieve optimal density with respect to any superlinear deadline function. We also show this is tight by ruling out timely generation with linear deadlines and vanishing hallucination rate.

Atul Ganju, Travis McVoy, Shaddin Dughmi +1

University of Southern California

Deadlines Inductive Bias Theory Preference Alignment Learning Higher-Order Structures

May 11, 2026cs.GT

When to Ask a Question: Understanding Communication Strategies in Generative AI Tools

Generative AI models differ from traditional machine learning tools in that they allow users to provide as much or as little information as they choose in their inputs. This flexibility often leads users to omit certain details, relying on the models to infer and fill in under-specified information based on distributional knowledge of user preferences. Such inferences may privilege majority viewpoints and disadvantage users with atypical preferences, raising concerns about fairness. Unlike more traditional recommender systems, LLMs can explicitly solicit more information from users through natural language. However, while directly eliciting user preferences could increase personalization and mitigate inequality, excessive querying places a burden on users who value efficiency. We develop a stylized model of user-LLM interaction and develop an objective that captures tradeoff between user burden and preference representation. Building on the observation that individual preferences are often correlated, we analyze how AI systems should balance inference and elicitation, characterizing the optimal amount of information to solicit before content generation. Ultimately, we show that information elicitation can mitigate the systematic biases of preference inference, enabling the design of generative tools that better incorporate diverse user perspectives while maintaining efficiency. We complement this theoretical analysis with an empirical evaluation illustrating the model's predictions and exploring their practical implications.

Charlotte Park, Kate Donahue, Manish Raghavan

Massachusetts Institute of Technology Cambridge, Massachusetts, USA

Generative Artificial Intelligence Elicitation Preference Alignment Learning Question

May 11, 2026cs.CL

DGPO: Beyond Pairwise Preferences with Directional Consistent Groupwise Optimization

Although Large Language Models (LLMs) have made remarkable progress, current preference optimization methods still struggle to align directional consistency while preserving reasoning diversity. To address this limitation, we propose Directional-Groupwise Preference Optimization (DGPO), a lightweight framework that aggregates supervision signals at the group level and explicitly models direction-aware alignment through multi-candidate comparisons. DGPO organizes forward and reverse question-answer instances into structured sets and optimizes a margin-based likelihood objective that separates coherent reasoning paths from inconsistent alternatives. This group-wise formulation captures richer relative information than pairwise objectives and reinforces consistency across diverse reasoning pathways. Empirical results show that our constructed reverse data yields a 3.2% average improvement across five benchmarks, while DGPO further delivers consistent gains across multiple datasets and model families, achieving average accuracy improvements of up to 3.6%.

Mengyi Deng, Zhiwei Li, Xin Li +4

1Information Hub, The Hong Kong University of Science and Technology (Guangzhou), China · 2The Hong Kong University of Science and Technology, Hong Kong SAR

Backtranslation Augmented Direct Preference Optimization Preference Alignment Learning LLM Reasoning Strategies Latent Representation Alignment

May 9, 2026cs.AI

Learning the Preferences of a Learning Agent

For AI systems to be useful to humans, they must understand and act in accordance with our values and preferences. Since specifying preferences is a hard task, inverse reinforcement learning (IRL) aims to develop methods that allow for inferring preferences from observed behavior. However, IRL assumes the human to be approximately optimal. This is a big limitation in cases where the human themselves may be learning to act optimally in an environment. In this paper, we formalize the problem of learning the preferences of a learning agent: a predictor observes a learner acting online and tries to infer the underlying reward function being (initially suboptimally) optimized by the learner. We model the learner as either being no-regret, or as converging to an optimal Boltzmann policy over time. In each of these settings, we establish theoretical guarantees for various preference learning algorithms, or otherwise show that such guarantees are impossible.

Karim Abdel Sadek, Mark Bedaywi, Rhys Gould +1

University of California, Berkeley

Preference Learning Reinforcement Learning From Human Feedback Preference Alignment Learning Offline Reinforcement Learning Boltzmann Policies Agentic Framework

May 9, 2026cs.LG

Personalized Alignment Revisited: The Necessity and Sufficiency of User Diversity

Personalized alignment aims to adapt large language models to heterogeneous user preferences, yet the precise theoretical conditions for its statistical efficiency have not been formally established. This paper characterizes the conditions under which personalized alignment achieves O(1) online regret and log(1/epsilon) offline sample complexity. We show that these optimal rates depend on a specific user-diversity condition: the population of user-specific heads must span the latent reward directions that can alter the optimal response. We prove that this condition is both necessary and sufficient. When it holds, simple greedy algorithms achieve benchmark efficiency; when it fails, every learner in a natural admissible class incurs at least logarithmic regret. Our results identify user diversity as the fundamental driver of personalized identifiability.

Enoch Hyunwook Kang

Foster School of Business, University of Washington, Seattle, Washington, USA.

Diversity Preference Alignment Learning Identifiability Sufficiency

May 9, 2026cs.LG

Mechanistic Analysis of Alignment Algorithms in Language Models

Post-training alignment algorithms are predominantly evaluated as black boxes, obscuring how they reshape language models' internal computations. We present a systematic mechanistic analysis of six preference-optimization methods: PPO, DPO, SimPO, ORPO, GRPO, and KTO across three open-weight model families. By integrating layer-wise linear probing, Sparse Autoencoders, and crosscoders, we localize preference representations and quantify alignment-induced geometric transformations in latent space. We find that preference signals consistently concentrate in early--mid or mid--late layers, but different objectives induce qualitatively distinct representational shifts. KTO and GRPO enhance linear separability through constructive feature sharing and sparse, high-salience recruitment. In contrast, DPO and ORPO degrade separability via non-constructive geometric rotation and feature attenuation, while PPO and SimPO largely preserve baseline geometry. These transformations exhibit architecture-dependent variability, demonstrating that behavioral alignment does not imply uniform internal restructuring. Our findings establish alignment as a heterogeneous intervention, motivate standardized feature-level auditing for safety and interpretability, and highlight the need for mechanism-aware optimization objectives.

Aarush Sinha, Ishan Garg, Veeraraju Elluru +2

University of Copenhagen · Independent · IIT Jodhpur +2

Large Language Model Alignment Preference Alignment Learning Representational Collapse Visual Geometry Grounded Transformer

May 8, 2026cs.LG

Can Revealed Preferences Clarify LLM Alignment and Steering?

LLMs are increasingly used to make or support high-stakes decisions under uncertainty, where alignment depends not only on factual accuracy but on how models weigh tradeoffs between different outcomes. We present an empirical pipeline for estimating the implied preferences that an LLM's observed choices optimize: we elicit the model's probability distribution over unknowns along with the choice it would make for the decision task and then fit a discrete choice model to recover the cost function that best rationalizes the model's decisions. We show how this revealed-preference description allows rigorous evaluation of whether models behave in a consistently goal-directed way, whether they can verbalize a description of their objectives which matches their revealed decision policy, and whether prompting can reliably steer those policies to implement a user-specified cost function. We apply this evaluation across four medical diagnosis domains and multiple frontier and open-source models. We find that while many models have a nontrivial degree of internal coherence, they also have significant weaknesses in faithfully reporting or adopting preferences in response to user direction.

Khurram Yamin, Jingjing Tang, Eric Horvitz +1

1Carnegie Mellon University · 2Microsoft Research

Large Language Model Alignment Large Language Model Decisions Preference Alignment Learning High-Stakes Domains Discrete Choice Models Steering

May 8, 2026cs.AI

Embeddings for Preferences, Not Semantics

Modern AI is opening the door to collective decision-making in which participants express their views as free-form text rather than voting on a fixed set of candidates. A natural idea is to embed these opinions in a vector space so that the substantial literature on facility location problems and fair clustering can be brought to bear. But standard text embeddings measure semantic similarity, whereas distances in facility location problems and fair clustering require what we call \textit{preferential similarity}: a participant's agreement with a piece of text should be inversely related to their distance from it. Off-the-shelf embeddings inherit a coarse preference signal through a correlation between semantic and preferential similarity, but fail to capture preferences when the correlation breaks. We formalize this as an invariance problem: text embedding models encode both a preference-relevant signal (stance and values) and semantic nuisance (style and wording), and the two are observationally correlated, so a geometry that relies on nuisance can appear preference-correct even when it is not. We show that synthetic training data designed to break this correlation provably shifts the optimal scorer away from nuisance-dominated cosine and significantly improves preference prediction across 11 online deliberation datasets.

Carter Blair, Ariel D. Procaccia, Milind Tambe

Harvard University

Preference Alignment Learning Text Embeddings Similarity and Metrics Deliberation

May 8, 2026cs.AI

Auto-Rubric as Reward: From Implicit Preferences to Explicit Multimodal Generative Criteria

Aligning multimodal generative models with human preferences demands reward signals that respect the compositional, multi-dimensional structure of human judgment. Prevailing RLHF approaches reduce this structure to scalar or pairwise labels, collapsing nuanced preferences into opaque parametric proxies and exposing vulnerabilities to reward hacking. While recent Rubrics-as-Reward (RaR) methods attempt to recover this structure through explicit criteria, generating rubrics that are simultaneously reliable, scalable, and data-efficient remains an open problem. We introduce Auto-Rubric as Reward (ARR), a framework that reframes reward modeling from implicit weight optimization to explicit, criteria-based decomposition. Before any pairwise comparison, ARR externalizes a VLM's internalized preference knowledge as prompt-specific rubrics, translating holistic intent into independently verifiable quality dimensions. This conversion of implicit preference structure into inspectable, interpretable constraints substantially suppresses evaluation biases including positional bias, enabling both zero-shot deployment and few-shot conditioning on minimal supervision. To extend these gains into generative training, we propose Rubric Policy Optimization (RPO), which distills ARR's structured multi-dimensional evaluation into a robust binary reward, replacing opaque scalar regression with rubric-conditioned preference decisions that stabilize policy gradients. On text-to-image generation and image editing benchmarks, ARR-RPO outperforms pairwise reward models and VLM judges, demonstrating that explicitly externalizing implicit preference knowledge into structured rubrics achieves more reliable, data-efficient multimodal alignment, revealing that the bottleneck is the absence of a factorized interface, not a deficit of knowledge.

Juanxi Tian, Fengyuan Liu, Jiaming Han +6

1Nanyang Technological University · 2Ant Group · 3MMLab, The Chinese University of Hong Kong +1

Progress Reward Modeling Task-Specific Rubrics Multimodal Generation Preference Alignment Learning Multi-Dimensional Evaluation

May 8, 2026cs.LG

Beyond Pairs: Your Language Model is Secretly Optimizing a Preference Graph

Direct Preference Optimization (DPO) aligns language models using pairwise preference comparisons, offering a simple and effective alternative to Reinforcement Learning (RL) from human feedback. However, in many practical settings, training data consists of multiple rollouts per prompt, inducing rich preference structure that pairwise DPO fails to exploit. Collapsing such data into independent pairs discards transitivity, introduces redundant or conflicting supervision, and can lead to unstable optimization. We propose Graph Direct Preference Optimization (GraphDPO), a principled generalization of DPO that operates over directed acyclic preference graphs induced by rollout rankings. GraphDPO encodes dominance relations as edges and optimizes a graph-structured Plackett--Luce-inspired objective that aggregates supervision over graph neighborhoods, enforcing transitivity while recovering standard DPO as a special case. To handle discrete or sparse signals, we introduce an equivalence-class construction where responses with identical preferences form graph layers, and intra-layer edges contribute zero loss, preventing spurious gradients. Despite leveraging full graph structure, GraphDPO maintains linear per-prompt complexity via efficient log-sum-exp aggregation. We further incorporate optional ground-truth anchoring by inserting verified solutions as dominant nodes and applying an annealed schedule that stabilizes early training while gradually relaxing oracle supervision. Experiments on reasoning and program synthesis tasks demonstrate superior performance, suggesting that graph-structured preference modeling is a scalable and robust alternative to pairwise and listwise alignment objectives.

Ning Liu, Chuanneng Sun, Kristina Klinkner +1

Amazon

Backtranslation Augmented Direct Preference Optimization Preference Alignment Learning Generalisation Beyond Retrieval

May 8, 2026cs.GT

Nash without Numbers: A Social Choice Approach to Mixed Equilibria in Context-Ordinal Games

Nash equilibrium serves as a fundamental mathematical tool in economics and game theory. However, it classically assumes knowledge of player utilities, whereas economics generally regards preferences as more fundamental. To leverage equilibrium analysis in strategic scenarios, one must first elicit numerical utilities consistent with player preferences, a delicate and time-consuming process. In this work, we forgo precise utilities and generalize the Nash equilibrium to a setting where we only assume a player is capable of providing an ordinal ranking of their actions within the context of other players' joint actions. The key technical challenge is to rethink the definition of a best-response. While the classical definition identifies actions maximizing expected payoff, we naturally look towards social choice theory for how to aggregate preferences to identify the most preferred actions. We define this generalized notion of a context-ordinal Nash equilibrium, establish its existence under mild conditions on aggregation methods, introduce notions of regularization, approximation, and regret, explore complexity for simple settings, and develop learning rules for computing such equilibria. In doing so, we provide a generalization of Nash equilibrium and demonstrate its direct applicability to elicited preferences in human experiments.

Ian Gemp, Crystal Qian, Marc Lanctot +1

Google DeepMind · New York, NY, USA · New York, NY USA +2

Nash Equilibrium Equilibrium Utility Maximization Preference Alignment Learning

May 8, 2026cs.CL

CLIPer: Tailoring Diverse User Preference via Classifier-Guided Inference-Time Personalization

Personalized LLMs can significantly enhance user experiences by tailoring responses to preferences such as helpfulness, conciseness, and humor. However, fine-tuning models to address all possible combinations of user preferences is computationally expensive and impractical. In this paper, we introduce \textbf{CLIPer}(\textbf{Cl}assifier-guided \textbf{I}nference-time \textbf{Per}sonalization), a lightweight personalization approach that leverages a classifier model to steer LLM generation dynamically to different user preferences at inference time. Our method eliminates the need for extensive fine-tuning, inducing negligible additional computational overhead while enabling more controllable and nuanced personalization across single and multi-dimensional preferences. Comprehensive empirical analyses demonstrate the scalability and effectiveness of our approach in delivering personalized language generation.

Jinyan Su, Jinpeng Zhou, Claire Cardie +1

Cornell University

Large Language Model Personalization Personalization Preference Alignment Learning Inference-Time

May 7, 2026cs.LG

Response Time Enhances Alignment with Heterogeneous Preferences

Aligning large language models (LLMs) to human preferences typically relies on aggregating pooled feedback into a single reward model. However, this standard approach assumes that all labelers share the same underlying preferences, ignoring the fact that real-world labelers are highly heterogeneous and usually anonymous. Consequently, relying solely on binary choice data fundamentally distorts the learned policy, making the true population-average preference unidentifiable. To overcome this critical limitation, we demonstrate that augmenting preference datasets with a simple, secondary signal -- the user's response time -- can restore the identifiability of the population's average preference. By modeling each decision as a Drift-Diffusion Model (DDM), we introduce a novel, consistent estimator of heterogeneous preferences that successfully corrects the distortions of standard choice-only labels. We prove that our estimator asymptotically converges to the true average preference even in extreme cases where each anonymous labeler contributes only a single choice. Empirically, across both synthetic and real-world datasets, our method consistently outperforms standard baselines that otherwise fail and plateau at a bias floor. Because response times are essentially free to record and require zero user tracking or identification, our results bring promises and open up new opportunities for future data-collection pipelines to improve the social benefit without requiring user-level identifiers or repeated elicitations.

Federico Echenique, Alireza Fallah, Baihe Huang +1

Department of Economics, University of California, Berkeley · Department of Computer Science and Ken Kennedy Institute, Rice University · Departments of Electrical Engineering and Computer Sciences, University of California, Berkeley +2

Preference Alignment Learning Large Language Model Personalization Reinforcement Learning From Human Feedback

May 7, 2026cs.LG

Preference Instability in Reward Models: Detection and Mitigation via Sparse Autoencoders

Preference learning in large language models relies on reward models as proxies for human judgment. However, these models frequently exhibit preference instability, producing contradictory preference assignments in response to subtle, meaning-preserving input variations. We analyze this instability at the representation level under three semantic-preserving perturbation types: paraphrasing, pattern injection, and backdoor triggers. We attribute this instability to over-reliance on predictive yet brittle features, which we term unstable features, and isolate them via Sparse Autoencoders (SAEs) in a sparse latent space where benign and perturbed inputs activate distinctly separable patterns. Building on this separability, we propose two SAE-based instability mitigation strategies: SAE Feature Steering, which identifies and suppresses anomalously activated features at inference, and SAE Residual Correction, which learns adaptive adjustments over SAE features to restore correct preferences. Our methods substantially reduce incorrect preference assignments on harmlessness and hallucination benchmarks while preserving benign performance and general utility on other tasks, without retraining the reward model.

Shunchang Liu, Xin Chen, Belen Martin Urcelay +1

ETH Zürich · Georgia Institute of Technology · ELLIS Institute Finland +1

Preference Learning Preference Alignment Learning Instability

May 6, 2026cs.CL

Misaligned by Reward: Socially Undesirable Preferences in LLMs

Reward models are a key component of large language model alignment, serving as proxies for human preferences during training. However, existing evaluations focus primarily on broad instruction-following benchmarks, providing limited insight into whether these models capture socially desirable preferences. As a result, important failures in social alignment can remain hidden. We extend reward-model benchmarking to four socially consequential domains: bias, safety, morality, and ethical reasoning. We introduce a framework that converts social evaluation datasets into pairwise preference data, leveraging gold labels where available and directional bias indicators otherwise. This enables us to test whether reward models prefer socially undesirable responses, and whether their preferences produce systematically biased distributions over selected outputs. Across five publicly available reward models and two instruction-tuned models used as reward proxies, we find substantial variation across domains, with no single model performing best overall. The models fall well short of strong social intelligence: they often prefer socially undesirable options, and their preferences produce systematically biased distributions. Moreover, stronger bias avoidance can reduce sensitivity to context, revealing a key alignment trade-off between avoiding biased outcomes and preserving contextual faithfulness. These findings show that standard reward benchmarks are insufficient for assessing social alignment and highlight the need for evaluations that directly measure the social preferences encoded in reward models.

Gayane Ghazaryan, Esra Dönmez

Institute for Natural Language Processing, University of Stuttgart · 2Interchange Forum for Reflecting on Intelligent Systems, University of Stuttgart

Large Language Model Alignment Large Language Model Bias Progress Reward Modeling Pluralistic Alignment Preference Alignment Learning Large Language Model Evaluation

May 6, 2026cs.CL

StoryAlign: Evaluating and Training Reward Models for Story Generation

Story generation aims to automatically produce coherent, structured, and engaging narratives. Although large language models (LLMs) have significantly advanced text generation, stories generated by LLMs still diverge from human-authored works regarding complex narrative structure and human-aligned preferences. A key reason is the absence of effective modeling of human story preferences, which are inherently subjective and under-explored. In this work, we systematically evaluate the modeling of human story preferences and introduce StoryRMB, the first benchmark for assessing reward models on story preferences. StoryRMB contains

1,133

high-quality, human-verified instances, each consisting of a prompt, one chosen story, and three rejected stories. We find existing reward models struggle to select human-preferred stories, with the best model achieving only

66.3\%

accuracy. To address this limitation, we construct roughly

100,000

high-quality story preference pairs across diverse domains and develop StoryReward, an advanced reward model for story preference trained on this dataset. StoryReward achieves state-of-the-art (SoTA) performance on StoryRMB, outperforming much larger models. We also adopt StoryReward in downstream test-time scaling applications for best-of-n (BoN) story selection and find that it generally chooses stories better aligned with human preferences. We will release our dataset, model, and code to facilitate future research. Related code and data are available at https://github.com/THU-KEG/StoryReward.

Haotian Xia, Hao Peng, Yunjia Qi +4

Department of Computer Science and Technology, Tsinghua University

Narrative Generation Narratives Preference Datasets Progress Reward Modeling Preference Alignment Learning Reward Functions

May 5, 2026cs.AI

Replacing Parameters with Preferences: Federated Alignment of Heterogeneous Vision-Language Models

Vision-Language Models (VLMs) have broad potential in privacy-sensitive domains such as healthcare and finance, yet strict data-sharing constraints render centralized training infeasible. Federated Learning mitigates this issue by enabling decentralized training, but practical deployments face challenges due to client heterogeneity in computational resources, application requirements, and model architectures. Under extreme model and data heterogeneity, replacing parameter aggregation with preference-based collaboration offers a more suitable interface, as it eliminates the need for direct parameter or data exchange. Motivated by this, we propose MoR, a federated alignment framework that combines GRPO with Mixture-of-Rewards for heterogeneous VLMs. In MoR, each client locally trains a reward model from local preference annotations, capturing specific evaluation signals without exposing raw data. To combine these heterogeneous supervision signals, MoR introduces a Mixture-of-Rewards mechanism with learned routing, which adaptively fuses client reward models according to the input and alignment objective. The server then optimizes a base VLM using GRPO with a KL penalty to a reference model, enabling preference alignment without requiring client models to share architectures or parameters. Experiments on diverse public vision-language benchmarks demonstrate that MoR consistently outperforms federated alignment baselines in generalization and cross-client adaptability. Our approach provides a scalable solution for privacy-preserving alignment of heterogeneous VLMs under federated settings.

Shule Lu, Yujing Wang, Hainan Zhang +5

Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing · Institute of Artificial Intelligence, Beihang University, China · MAIS, Institute of Automation, Chinese Academy of Sciences, China +2

Vision-Language Alignment Recent Vision-Language Models Preference Alignment Learning

May 4, 2026cs.AI

U-Define: Designing User Workflows for Hard and Soft Constraints in LLM-Based Planning

LLMs are increasingly used for end-user task planning, yet their black-box nature limits users' ability to ensure reliability and control. While recent systems incorporate verification techniques, it remains unclear how users can effectively apply such rigid constraints to represent intent or adapt to real-world variability. For example, prior work finds that hard-only constraints are too rigid, and numeric flexibility weights confuse users. We investigate how interaction workflows can better support users in applying constraints to guide LLM-generated plans, examining whether abstracting strictness into high-level types (i.e., hard and soft) paired with distinct verification mechanisms helps users more reliably express and align intent. We present U-Define, a system that lets users define constraints in natural language and categorize them as either hard rules that must not be violated or soft preferences that allow flexibility. U-Define verifies these types through complementary methods: formal model checking for hard constraints and LLM-as-judge evaluation for soft ones. Through a technical evaluation and user studies with general and expert participants, we find that user-defined constraint types improve perceived usefulness, performance, and satisfaction while maintaining usability. These findings provide insights for designing flexible yet reliable constraint-based workflows.

Christine P Lee, Xinyu Jessica Wang, Aws Albarghouthi +2

University of Wisconsin–Madison, USA · George Mason University, USA

Agentic Workflow Design Structural Constraints Planning Interaction Data Preference Alignment Learning

May 3, 2026cs.LG

Multi-User Dueling Bandits: A Fair Approach using Nash Social Welfare

Learning from human preference data is becoming a useful tool, from fine-tuning large language models to training reinforcement learning agents. However, in most scenarios, the model is trained on the average preference of all human evaluators, which, under large variations of preferences, can be unfair to minority groups. In this work, we consider fairness in dueling bandits, a standard framework for online learning from preference data. We assume that each user has a (potentially distinct) Condorcet winner, which is an arm preferred to every other arm. Using these user-specific Condorcet winners as reference points, we evaluate and score arms according to their performance relative to the corresponding winner. To promote fairness across heterogeneous users, we adopt the well-established Nash Social Welfare objective, which maximizes the product of user utilities, thereby inherently penalizing inequality and preventing the marginalization of any single user. Within this framework, we construct a hard instance to establish a regret lower bound of

Ω(T^{2/3}\min(K,D)^\frac{1}{3})

for a time horizon

T

,

K

arms, and

D

users, which, to the best of our knowledge, is the first result quantifying the cost of fairness in dueling bandits with heterogeneous preferences. We then present the Fair-Explore-Then-Commit and Fair-

ε

-Greedy algorithms with a Condorcet winner identification phase. We further derive their regret upper bounds that match the lower-bound dependence on

T

up to logarithmic factors.

Maheed H. Ahmed, Mahsa Ghasemi

Electrical and Computer Engineering Purdue University

Algorithmic Fairness Stochastic Multi-Armed Bandits Bandits Preference Alignment Learning Greedy Algorithm

May 3, 2026cs.CL

RMGAP: Benchmarking the Generalization of Reward Models across Diverse Preferences

Reinforcement Learning from Human Feedback has become the standard paradigm for language model alignment, where reward models directly determine alignment effectiveness. In this work, we focus on how to evaluate the generalizability of reward models. By "generalizability", we mean the ability of RMs to correctly rank responses to align with diverse user preferences. However, existing reward model benchmarks are typically designed around a universal preference, failing to assess this generalization. To address this critical gap, we introduce RMGAP, a benchmark comprising 1,097 instances across Chat, Writing, Reasoning, and Safety domains. Since different users exhibit diverse preferences for the same task, we first generate four distinct responses with different linguistic profiles for each collected prompt. However, the original prompt set lacks the specificity to convey different preferences. We therefore construct tailored prompts by contrasting these candidates and designing scenarios in which one response becomes the uniquely appropriate choice. Moreover, we observe that users often express the same preference using different phrasings, and thus extend each prompt with two paraphrased variants. Our evaluation of 24 state-of-the-art RMs reveals their substantial limitations: even the best RM achieves only 49.27% Best-of-N accuracy, highlighting considerable room for improvement in reward model generalization. Related data and code are available at https://github.com/nanzhi84/RMGAP.

Yangyang Zhou, Yi-Chen Li

Beijing University of Posts and Telecommunications · National Key Laboratory for Novel Software Technology, Nanjing University, China · School of Artificial Intelligence, Nanjing University, China

Progress Reward Modeling Large Language Model Alignment Preference Alignment Learning

May 2, 2026cs.RO

VLA-ATTC: Adaptive Test-Time Compute for VLA Models with Relative Action Critic Model

Vision-Language-Action (VLA) models have demonstrated remarkable capabilities and generalization in embodied manipulation. However, their decision-making relies on a fast, instinctive process that lacks deliberation. This strategy often leads to suboptimal or catastrophic actions when facing complex or ambiguous scenarios that require greater consideration. In this paper, we introduce \textbf{VLA-ATTC}, a framework that endows VLA models with adaptive test-time compute (TTC). VLA-ATTC employs an uncertainty-based ``cognitive clutch'' to dynamically transition from reflexive execution to a TTC deliberation phase when necessary. During TTC phase, a novel \textbf{Relative Action Critic} (RAC) model identifies the optimal action from generated candidates via pairwise comparisons. This relative mechanism replaces unstable absolute value estimation, significantly simplifying the learning objective. Furthermore, we introduce an efficient sampling strategy to amortize computational costs and an automated data pipeline that curates preference pairs without manual annotation. On the LIBERO-LONG benchmark, VLA-ATTC reduces the failure rate of the SOTA model PI0.5 by over 50%. We will open-source all the code and weights.

Wenhao Li, Xiu Su, Yichao Cao +5

University of Sydney · Central South University · University of Science and Technology of China +2

Diffusion-Based Vision-Language-Actions Critic-Free Reinforcement Learning Test Time Cognitive Science Preference Alignment Learning

Apr 30, 2026cs.LG

Mind the Gap: Structure-Aware Consistency in Preference Learning

Preference learning has become the foundation of aligning Large Language Models (LLMs) with human intent. Popular methods, such as Direct Preference Optimization (DPO), minimize surrogate losses as proxies for the intractable pairwise ranking loss. However, we demonstrate that for the equicontinuous hypothesis sets typical of neural networks, these standard surrogates are theoretically inconsistent, yielding vacuous generalization guarantees. To resolve this, we formulate LLM alignment within a margin-shifted ranking framework. We derive rigorous

H

-consistency bounds that depend on enforcing a separation margin

γ

. Crucially, we extend this to Structure-Aware

H

-consistency, introducing a novel objective (SA-DPO) that adapts the margin based on the semantic distance between responses to handle synonyms and hard pairs. Finally, we analyze the trade-off between consistency and model limitations via the Margin-Capacity Profile, proving that heavy-tailed surrogates (such as the Polynomial Hinge family) offer superior consistency guarantees for capacity-bounded models compared to the standard logistic loss used in DPO.

Mehryar Mohri, Yutao Zhong

Google Research, New York, NY · Courant Institute of Mathematical Sciences, New York, NY

Large Language Model Alignment Preference Alignment Learning Structure-Aware Upper Bounds

Apr 29, 2026cs.MA

When Agents Shop for You: Role Coherence in AI-Mediated Markets

Consumers are increasingly delegating purchase decisions to AI agents, providing natural-language descriptions of their preferences and identity. We argue that these representations constitute an information channel, role coherence, through which sellers can infer willingness to pay without explicit disclosure by the buyer agent, leading to preference leakage. In an experiment where a language-model buyer agent shops on behalf of a verbal consumer profile, we show that seller-side inference from dialogue alone recovers willingness to pay nearly one-for-one. Comparing this setting to a numeric-budget condition with confidentiality instructions cleanly isolates role coherence as distinct from instruction-following failure. Because this leakage arises from delegation itself, it cannot be mitigated at the prompt level. Instead, we propose architectural interventions that trade off personalization against preference privacy.

Soogand Alavi, Salar Nozari

Tippie College of Business, University of Iowa

Agentic Commerce Preference Alignment Learning Artificial Intelligence Agents Role Profile

Apr 24, 2026cs.AI

On the Existence of an Inverse Solution for Preference-Based Reductions in Argumentation

Preference-based argumentation frameworks (PAFs) extend Dung's approach to abstract argumentation (AAFs) by encoding preferences over arguments. Such preferences control the transformation of attacks into defeats, and different approaches to doing so result in different reductions from a PAF to an AAF. In this paper we consider a PAF inverse problem which takes an argumentation graph, a labelling and a semantics as an input, and outputs a yes" or no" as to whether there is a preference relation between the arguments which can yield the desired labelling. This inverse problem has applications in areas including preference elicitation and explainability. We consider this problem in the context of the four most widely-used preference based reductions under the complete semantics. We show that in most cases, the problem can be answered in polynomial time.

Alessio Zaninotto, Bruno Yun, Nir Oren +1

1CRIL - Université d’Artois, CNRS · 2Universite Claude Bernard Lyon 1, CNRS, Ecole Centrale de Lyon, INSA Lyon, Université Lumière Lyon 2, LIRIS, UMR5205, 69622 Villeurbanne, France · University of Aberdeen

Argumentation Frameworks Preference Alignment Learning Inverse Design

Apr 24, 2026cs.CL

Preference Heads in Large Language Models: A Mechanistic Framework for Interpretable Personalization

Large Language Models (LLMs) exhibit strong implicit personalization ability, yet most existing approaches treat this behavior as a black box, relying on prompt engineering or fine tuning on user data. In this work, we adopt a mechanistic interpretability perspective and hypothesize the existence of a sparse set of Preference Heads, attention heads that encode user specific stylistic and topical preferences and exert a causal influence on generation. We introduce Differential Preference Steering (DPS), a training free framework that (1) identifies Preference Heads through causal masking analysis and (2) leverages them for controllable and interpretable personalization at inference time. DPS computes a Preference Contribution Score (PCS) for each attention head, directly measuring its causal impact on user aligned outputs. During decoding, we contrast model predictions with and without Preference Heads, amplifying the difference between personalized and generic logits to selectively strengthen preference aligned continuations. Experiments on widely used personalization benchmarks across multiple LLMs demonstrate consistent gains in personalization fidelity while preserving content coherence and low computational overhead. Beyond empirical improvements, DPS provides a mechanistic explanation of where and how personalization emerges within transformer architectures. Our implementation is publicly available.

Weixu Zhang, Ye Yuan, Changjiang Han +7

McGill University · Mila - Quebec AI Institute · MBZUAI +2

Large Language Model Personalization Personalization Preference Alignment Learning Steering

Apr 23, 2026cs.CL

Preferences of a Voice-First Nation: Large-Scale Pairwise Evaluation and Preference Analysis for TTS in Indian Languages

Crowdsourced pairwise evaluation has emerged as a scalable approach for assessing foundation models. However, applying it to Text to Speech(TTS) introduces high variance due to linguistic diversity and multidimensional nature of speech perception. We present a controlled multidimensional pairwise evaluation framework for multilingual TTS that combines linguistic control with perceptually grounded annotation. Using 5K+ native and code-mixed sentences across 10 Indic languages, we evaluate 7 state-of-the-art TTS systems and collect over 120K pairwise comparisons from over 1900 native raters. In addition to overall preference, raters provide judgments across 6 perceptual dimensions: intelligibility, expressiveness, voice quality, liveliness, noise, and hallucinations. Using Bradley-Terry modeling, we construct a multilingual leaderboard, interpret human preference using SHAP analysis and analyze leaderboard reliability alongside model strengths and trade-offs across perceptual dimensions.

Srija Anand, Ashwin Sankar, Ishvinder Sethi +10

Indian Institute of Technology Madras, India · AI4Bharat, India · Josh Talks, India

Seed-Tts-Eval Benchmark Multilingual Automatic Speech Recognition Speaker Linguistics Indian Languages Preference Alignment Learning

Apr 23, 2026cs.AI

Align Generative Artificial Intelligence with Human Preferences: A Novel Large Language Model Fine-Tuning Method for Online Review Management

Online reviews have played a pivotal role in consumers' decision-making processes. Existing research has highlighted the significant impact of managerial review responses on customer relationship management and firm performance. However, a large portion of online reviews remains unaddressed due to the considerable human labor required to respond to the rapid growth of online reviews. While generative AI has achieved remarkable success in a range of tasks, they are general-purpose models and may not align well with domain-specific human preferences. To tailor these general generative AI models to domain-specific applications, finetuning is commonly employed. Nevertheless, several challenges persist in finetuning with domain-specific data, including hallucinations, difficulty in representing domain-specific human preferences, and over conservatism in offline policy optimization. To address these challenges, we propose a novel preference finetuning method to align an LLM with domain-specific human preferences for generating online review responses. Specifically, we first identify the source of hallucination and propose an effective context augmentation approach to mitigate the LLM hallucination. To represent human preferences, we propose a novel theory-driven preference finetuning approach that automatically constructs human preference pairs in the online review domain. Additionally, we propose a curriculum learning approach to further enhance preference finetuning. To overcome the challenge of over conservatism in existing offline preference finetuning method, we propose a novel density estimation-based support constraint method to relax the conservatism, and we mathematically prove its superior theoretical guarantees. Extensive evaluations substantiate the superiority of our proposed preference finetuning method.

Yanan Wang, Yong Ge

Department of Information Systems and Operations Management, The University of Texas at Arlington · Department of Management Information Systems, University of Arizona

Preference Alignment Learning Review Large Language Model Fine-Tuning Reinforcement Learning From Human Feedback Model Fine-Tuning Generative Artificial Intelligence+1

Apr 22, 2026cs.AI

Where and What: Reasoning Dynamic and Implicit Preferences in Situated Conversational Recommendation

Situated conversational recommendation (SCR), which utilizes visual scenes grounded in specific environments and natural language dialogue to deliver contextually appropriate recommendations, has emerged as a promising research direction due to its close alignment with real-world scenarios. Compared to traditional recommendations, SCR requires a deeper understanding of dynamic and implicit user preferences, as the surrounding scene often influences users' underlying interests, while both may evolve across conversations. This complexity significantly impacts the timing and relevance of recommendations. To address this, we propose situated preference reasoning (SiPeR), a novel framework that integrates two core mechanisms: (1) Scene transition estimation, which estimates whether the current scene satisfies user needs, and guides the user toward a more suitable scene when necessary; and (2) Bayesian inverse inference, which leverages the likelihood of multimodal large language models (MLLMs) to predict user preferences about candidate items within the scene. Extensive experiments on two representative benchmarks demonstrate SiPeR's superiority in both recommendation accuracy and response generation quality. The code and data are available at https://github.com/DongdingLin/SiPeR.

Dongding Lin, Jian Wang, Yongqi Li +1

Department of Computing, The Hong Kong Polytechnic University · College of Computer Science, Sichuan University

Preference Alignment Learning Multimodal Large Language Models Generative Quality

Apr 22, 2026cs.HC

Vibrotactile Preference Learning: Uncertainty-Aware Preference Learning for Personalized Vibration Feedback

Individual differences in vibrotactile perception underscore the growing importance of personalization as haptic feedback becomes more prevalent in interactive systems. We propose Vibrotactile Preference Learning (VPL), a system that captures user-specific preference spaces over vibrotactile parameters via Gaussian-process-based uncertainty-aware preference learning. VPL uses an expected information gain-based acquisition strategy to guide query selection over 40 rounds of pairwise comparisons of overall user preference, augmented with user-reported uncertainty, enabling efficient exploration of the parameter space. We evaluate VPL in a user study (N = 13) using the vibrotactile feedback from a Microsoft Xbox controller, showing that it efficiently learns individualized preferences while maintaining comfortable, low-workload user interactions. These results highlight the potential of VPL for scalable personalization of vibrotactile experiences.

Rongtao Zhang, Xin Zhu, Masoume Pourebadi Khotbehsara +3

Department of Computer Science University of Southern California Los Angeles, California, USA

Preference Learning Haptic Feedback Preference Alignment Learning Vibration Personalization

Apr 22, 2026cs.AI

HiPO: Hierarchical Preference Optimization for Adaptive Reasoning in LLMs

Direct Preference Optimization (DPO) is an effective framework for aligning large language models with human preferences, but it struggles with complex reasoning tasks. DPO optimizes for the likelihood of generating preferred over dispreferred responses in their entirety and lacks the granularity to provide feedback on subsections of many-step solutions typical of reasoning tasks. Existing methods excel at either stable preference learning (e.g., DPO variants like KTO and RSO) or structured reasoning (e.g., ReMA's multi-agent RL framework, Tree of Thoughts), but fail to merge these complementary strengths. We propose HiPO (Hierarchical Preference Optimization), an extension of DPO that separates responses into reasoning segments (query clarification and context, reasoning steps, and answer) and computes loss as a weighted sum of the DPO loss for each segment. Our approach enables segment-specific training while maintaining DPO's computational efficiency and training stability. We demonstrate that for multiple 7B LLMs fine-tuned using HiPO and DPO on the Math Stack Exchange preference dataset, the models trained with HiPO outperform the others on a variety of common math benchmarks and achieve greater organization, logical flow, and consistency as measured by GPT-4.1.

Darsh Kachroo, Arjun Prasaath Anbazhagan, Adriana Caraeni +2

Vellore Institute of Technology, Chennai · Northwestern University · University of Massachusetts Amherst +2

Sequential Preference Optimization LLM Reasoning Strategies Reasoning Skills Preference Alignment Learning

Apr 21, 2026cs.AI

DT2IT-MRM: Debiased Preference Construction and Iterative Training for Multimodal Reward Modeling

Multimodal reward models (MRMs) play a crucial role in aligning Multimodal Large Language Models (MLLMs) with human preferences. Training a good MRM requires high-quality multimodal preference data. However, existing preference datasets face three key challenges: lack of granularity in preference strength, textual style bias, and unreliable preference signals. Besides, existing open-source multimodal preference datasets suffer from substantial noise, yet there is a lack of effective and scalable curation methods to enhance their quality. To address these limitations, we propose \textbf{DT2IT-MRM}, which integrates a \textbf{D}ebiased preference construction pipeline, a novel reformulation of text-to-image (\textbf{T2I}) preference data, and an \textbf{I}terative \textbf{T}raining framework that curates existing multimodal preference datasets for \textbf{M}ultimodal \textbf{R}eward \textbf{M}odeling. Our experimental results show that DT2IT-MRM achieves new \textbf{state-of-the-art} overall performance on three major benchmarks: VL-RewardBench, Multimodal RewardBench, and MM-RLHF-RewardBench.

Zhihong Zhang, Jie Zhao, Xiaojian Huang +5

University of Science and Technology of China

Multimodal Large Language Models Multimodal Dataset Preference Alignment Learning

Apr 21, 2026cs.AI

From Experience to Skill: Multi-Agent Generative Engine Optimization via Reusable Strategy Learning

Generative engines (GEs) are reshaping information access by replacing ranked links with citation-grounded answers, yet current Generative Engine Optimization (GEO) methods optimize each instance in isolation, unable to accumulate or transfer effective strategies across tasks and engines. We reframe GEO as a strategy learning problem and propose MAGEO, a multi-agent framework in which coordinated planning, editing, and fidelity-aware evaluation serve as the execution layer, while validated editing patterns are progressively distilled into reusable, engine-specific optimization skills. To enable controlled assessment, we introduce a Twin Branch Evaluation Protocol for causal attribution of content edits and DSV-CF, a dual-axis metric that unifies semantic visibility with attribution accuracy. We further release MSME-GEO-Bench, a multi-scenario, multi-engine benchmark grounded in real-world queries. Experiments on three mainstream engines show that MAGEO substantially outperforms heuristic baselines in both visibility and citation fidelity, with ablations confirming that engine-specific preference modeling and strategy reuse are central to these gains, suggesting a scalable learning-driven paradigm for trustworthy GEO. Code is available at https://github.com/Wu-beining/MAGEO

Beining Wu, Fuyou Mao, Jiong Lin +7

Hangzhou Dianzi University · Central South University · Shanghai Artificial Intelligence Laboratory +3

Agentic Optimization Citations Skills Preference Alignment Learning

Apr 21, 2026cs.AI

Personalized Benchmarking: Evaluating LLMs by Individual Preferences

With the rise in capabilities of large language models (LLMs) and their deployment in real-world tasks, evaluating LLM alignment with human preferences has become an important challenge. Current benchmarks average preferences across all users to compute aggregate ratings, overlooking individual user preferences when establishing model rankings. Since users have varying preferences in different contexts, we call for personalized LLM benchmarks that rank models according to individual needs. We compute personalized model rankings using ELO ratings and Bradley-Terry coefficients for 115 active Chatbot Arena users and analyze how user query characteristics (topics and writing style) relate to LLM ranking variations. We demonstrate that individual rankings of LLM models diverge dramatically from aggregate LLM rankings, with Bradley-Terry correlations averaging only

ρ= 0.04

(57% of users show near-zero or negative correlation) and ELO ratings showing moderate correlation (

ρ= 0.43

). Through topic modeling and style analysis, we find users exhibit substantial heterogeneity in topical interests and communication styles, influencing their model preferences. We further show that a compact combination of topic and style features provides a useful feature space for predicting user-specific model rankings. Our results provide strong quantitative evidence that aggregate benchmarks fail to capture individual preferences for most users, and highlight the importance of developing personalized benchmarks that rank LLM models according to individual user preferences.

Cristina Garbacea, Heran Wang, Chenhao Tan

University of Chicago

Large Language Model Personalization Preference Alignment Learning Chatbot Arena Mean

Apr 19, 2026cs.CL

HorizonBench: Long-Horizon Personalization with Evolving Preferences

User preferences evolve across months of interaction, and tracking them requires inferring when a stated preference has been changed by a subsequent life event. We define this problem as long-horizon personalization and observe that progress on it is limited by data availability and measurement, with no existing resource providing both naturalistic long-horizon interactions and the ground-truth provenance needed to diagnose why models fail. We introduce a data generator that produces conversations from a structured mental state graph, yielding ground-truth provenance for every preference change across 6-month timelines, and from it construct HorizonBench, a benchmark of 4,245 items from 360 simulated users with 6-month conversation histories averaging ~4,300 turns and ~163K tokens. HorizonBench provides a testbed for long-context modeling, memory-augmented architectures, theory-of-mind reasoning, and user modeling. Across 25 frontier models, the best model reaches 52.8% and most score at or below the 20% chance baseline. When these models err on evolved preferences, over a third of the time they select the user's originally stated value without tracking the updated user state. This belief-update failure persists across context lengths and expression explicitness levels, identifying state-tracking capability as the primary bottleneck for long-horizon personalization.

Shuyue Stella Li, Bhargavi Paranjape, Kerem Oktar +9

University of Washington · Meta · OpenAI +1

Preference Alignment Learning Interaction History Dialogue Benchmarks Horizon

Apr 19, 2026cs.CL

Beyond Overlap Metrics: Rewarding Reasoning and Preferences for Faithful Multi-Role Dialogue Summarization

Multi-role dialogue summarization requires modeling complex interactions among multiple speakers while preserving role-specific information and factual consistency. However, most existing methods optimize for automatic metrics such as ROUGE and BERTScore, which favor surface-level imitation of references rather than genuine gains in faithfulness or alignment with human preferences. We propose a novel framework that couples explicit cognitive-style reasoning with reward-based optimization for multi-role dialogue summarization. Our method first distills structured reasoning traces (e.g., step-by-step inferences and intermediate reflections) from a large teacher model and uses them as auxiliary supervision to initialize a reasoning-aware summarizer via staged supervised fine-tuning. It then applies GRPO with a dual-principle reward that blends metric-based signals with human-aligned criteria targeting key information coverage, implicit inference, factual faithfulness, and conciseness. Experiments on multilingual multi-role dialogue benchmarks show that our method matches strong baselines on ROUGE and BERTScore. Specifically, results on CSDS confirm the framework's stability in semantic consistency, while in-depth analysis on SAMSum demonstrates clear gains in factual faithfulness and model-based preference alignment. These findings underscore the value of reasoning-aware and preference-aware training for reliable dialogue summarization. Checkpoints and datasets are available at https://huggingface.co/collections/NebulaPixel/summorchestra-multirole-summary.

Xiaoyong Mei, Tingting Zuo, Da Chen +5

Zhejiang Normal University · Huawei Technologies · HKUST, Hong Kong +2

Dialogue Benchmarks Summarization Theory-Of-Mind Reasoning Preference Alignment Learning

Apr 18, 2026cs.CV

Comparison Drives Preference: Reference-Aware Modeling for AI-Generated Video Quality Assessment

The rapid advancement of generative models has led to a growing volume of AI-generated videos, making the automatic quality assessment of such videos increasingly important. Existing AI-generated content video quality assessment (AIGC-VQA) methods typically estimate visual quality by analyzing each video independently, ignoring potential relationships among videos. In this work, we revisit AIGC-VQA from an inter-video perspective and formulate it as a reference-aware evaluation problem. Through this formulation, quality assessment is guided not only by intrinsic video characteristics but also by comparisons with related videos, which is more consistent with human perception. To validate its effectiveness, we propose Reference-aware Video Quality Assessment (RefVQA), which utilizes a query-centered reference graph to organize semantically related samples and performs graph-guided difference aggregation from the reference nodes to the query node. Experiments on existing datasets demonstrate that our proposed RefVQA outperforms state-of-the-art methods across multiple quality dimensions, with strong generalization ability validated by cross-dataset evaluation. These results highlight the effectiveness of the proposed reference-based formulation and suggest its potential to advance AIGC-VQA.

Minghao Zou, Gen Liu, Guanghui Yue +5

Cardiff University Cardiff, UK · Shandong University of Science and Technology Qingdao, China · Shenzhen University Shenzhen, China +2

Video Quality Generative Video Models Cross-Dataset Benchmark Preference Alignment Learning

Apr 18, 2026cs.IR

From "Strings" to "Things" for Personal Knowledge Graphs: Evaluating LLM Triple Extraction for Recommendation Systems

Personal Knowledge Graphs (PKGs) offer a privacy-preserving framework for modeling user preferences, yet constructing them from unstructured, decentralized conversational data remains a challenge. This paper bridges the gap between conversational "strings" and semantic "things" by presenting a reproducible pipeline for extracting structured user-preference triples using lightweight Large Language Models. We evaluate Qwen- and Gemma-based models on their ability to extract RDF-compliant triples linked to Wikidata identifiers from conversational data for PKG construction. Our evaluation assesses both the semantic extraction fidelity and the utility of the resulting graphs in a downstream recommendation task. We found that certain models performed well and had proportionally high downstream performance relative to their triple extraction performance.

Abhirup Dasgupta, Fernando Spadea, Oshani Seneviratne

Rensselaer Polytechnic Institute, Troy NY 12180, USA

Rdf Knowledge Graphs Preference Alignment Learning

Apr 17, 2026cs.CL

Preference Estimation via Opponent Modeling in Multi-Agent Negotiation

Automated negotiation in complex, multi-party and multi-issue settings critically depends on accurate opponent modeling. However, conventional numerical-only approaches fail to capture the qualitative information embedded in natural language interactions, resulting in unstable and incomplete preference estimation. Although Large Language Models (LLMs) enable rich semantic understanding of utterances, it remains challenging to quantitatively incorporate such information into a consistent opponent modeling. To tackle this issue, we propose a novel preference estimation method integrating natural language information into a structured Bayesian opponent modeling framework. Our approach leverages LLMs to extract qualitative cues from utterances and converts them into probabilistic formats for dynamic belief tracking. Experimental results on a multi-party benchmark demonstrate that our framework improves the full agreement rate and preference estimation accuracy by integrating probabilistic reasoning with natural language understanding.

Yuta Konishi, Kento Yamamoto, Eisuke Sonomoto +8

1Graduate School of Informatics, Kyoto University, Kyoto, Japan · 2Accenture Japan Ltd, Tokyo, Japan

Negotiation Preference Alignment Learning Natural Language Processing Probabilistic Model

Apr 16, 2026cs.AI

WavAlign: Enhancing Intelligence and Expressiveness in Spoken Dialogue Models via Adaptive Hybrid Post-Training

End-to-end spoken dialogue models have garnered significant attention because they offer a higher potential ceiling in expressiveness and perceptual ability than cascaded systems. However, the intelligence and expressiveness of current open-source spoken dialogue models often remain below expectations. Motivated by the success of online reinforcement learning(RL) in other domains, one might attempt to directly apply preference optimization to spoken dialogue models, yet this transfer is non-trivial. We analyze these obstacles from the perspectives of reward modeling and rollout sampling, focusing on how sparse preference supervision interacts with dense speech generation under shared-parameter updates. Based on the analysis, we propose a modality-aware adaptive post-training recipe that makes RL practical for spoken dialogue: it constrains preference updates to the semantic channel and improves acoustic behavior via explicit anchoring, while dynamically regulating their mixture from rollout statistics to avoid unreliable preference gradients. We evaluate the method across multiple spoken dialogue benchmarks and representative architectures, and observe consistent improvements in semantic quality and speech expressiveness.

Yifu Chen, Shengpeng Ji, Qian Chen +9

Zhejiang University · Tongyi Fun Team, Alibaba Group · Beijing University of Technology

Speech Language Models Preference Alignment Learning Expressivity Offline Reinforcement Learning Intelligence

Apr 16, 2026cs.CL

CoPA: Benchmarking Personalized Question Answering with Data-Informed Cognitive Factors

While LLMs have demonstrated remarkable potential in Question Answering (QA), evaluating personalization remains a critical bottleneck. Existing paradigms predominantly rely on lexical-level similarity or manual heuristics, often lacking sufficient data-driven validation. We address this by mining Community-Individual Preference Divergence (CIPD), where individual choices override consensus, to distill six key personalization factors as evaluative dimensions. Accordingly, we introduce CoPA, a benchmark with 1,985 user profiles for fine-grained, factor-level assessment. By quantifying the alignment between model outputs and user-specific cognitive preferences inferred from interaction patterns, CoPA provides a more comprehensive and discriminative standard for evaluating personalized QA than generic metrics. The code is available at https://github.com/bjzgcai/CoPA.

Hang Su, Zequn Liu, Chen Hu +3

East China Normal University, China · Zhongguancun Academy, China

Large Language Model Personalization Preference Alignment Learning Cognitive Science Factor

Apr 14, 2026cs.AI

Rater State Bias in RLHF Preference Data: An Audit Framework

We identify a structured confound in Reinforcement Learning from Human Feedback (RLHF). Pairwise preference labels are intended to reflect the compared outputs, but they may also reflect the rater's state during annotation. Under sustained stressful or distressing conditions, raters' preferences may shift over time, so that preference data encode rater state alongside judgments about response quality. We argue that, if present, such shifts would differ from ordinary disagreement or random label noise. They would be state dependent, could be shared across annotators under similar conditions, and would not necessarily cancel during aggregation, reward modeling, and policy optimization. We propose rater state shift as a plausible and testable source of structured bias in RLHF preference data. This paper develops a hypothesis and an audit framework for studying this source of bias. We define rater state shift, rater state confound, and correlated rater state bias. We also propose survival level emotional authenticity as a candidate output signature, defined by lexical, pragmatic, discourse, and safety features whose reliability and validity remain to be demonstrated. We show that systematic rater state bias can survive aggregation and may enter the learned reward signal. We state five testable predictions, together with effect size thresholds for an initial audit, and note which require proprietary data. Finally, we present an audit protocol and pilot study plan that can be applied to publicly available instruction tuned models. We do not infer the training history of any specific deployed model.

Elena Kopteva, Vitaliy Hlynianyi-Zhuk

The Grainger College of Engineering, Department of Physics & Illinois Center for Advanced Studies of the Universe, University of Illinois Urbana-Champaign, Urbana, Illinois 61801, USA · Faculty of Applied Mathematics, Oles Honchar Dnipro National University, Dnipro, 49045, Ukraine · Department of Clinical Psychology, Kyiv Institute of Modern Psychology and Psychotherapy, Kyiv, 01133, Ukraine

Reinforcement Learning From Human Feedback Rater Preference Alignment Learning Model Auditing Reward Signal

Apr 10, 2026cs.LG

Relational Preference Encoding in Looped Transformer Internal States

We investigate how looped transformers encode human preference, training lightweight evaluator heads on frozen Ouro-2.6B loop-iteration states on Anthropic HH-RLHF. v2: an erratum is prepended; the original manuscript is unchanged. A post-publication audit found the three headline results inflated by two independent evaluation errors. The 95.2% pairwise evaluator accuracy is a canonical-ordering artifact: the data were correctly split, but the evaluator learned to prefer the first-presented argument; its strict antisymmetrized accuracy on the full 8,552-pair test set is 63.9%. The 84.5% pairwise probe and the below-chance 21.75% pointwise probe were source-item leaks (orientation rows and pair partners crossing the train/test split); corrected pair-disjoint values are 56.5% and 54.2% -- above chance, so the "inverted polarity" finding is withdrawn. The central finding survives at much smaller magnitude: preference is decoded more accurately relationally than pointwise (paired +2.3 points, 95% CI [+1.3, +3.3]), and the antisymmetrized evaluator still beats the linear probe, but no corrected readout rivals end-to-end reward models. The methodological findings stand (constant-output degeneracy, flip test, swap-protocol metric deflation), with one correction: antisymmetrized accuracy, not antisymmetry correlation, certifies relational discrimination. The two errors are mutually invisible -- a split audit cannot see an ordering prior, antisymmetrization cannot see a leak -- so both checks are required. Full audit in the follow-up work (Kirin, 2026, in preparation).

Jan Kirin

Preference Alignment Learning Transformer Architectures Progress Reward Modeling

Mar 11, 2026cs.CL

Coverage-Controlled Preference Mining from Noisy Claim Verification for Evidence-Grounded Generation

Evidence-grounded generation produces summaries whose claims should be supported by supplied evidence, but claim-level verifiers provide noisy feedback and can reward models that simply say less. We study this problem in clinical Brief Hospital Course summarization, where outputs must remain grounded in patient-specific EHR evidence. We introduce VERI-DPO, a preference-mining framework that converts noisy claim verification into coverage-controlled summary-level preferences. For each evidence-window prompt, VERI-DPO samples multiple candidate summaries, decomposes them into claims, verifies each claim against patient evidence, and forms a preference pair only when the chosen summary has better aggregate verifier-estimated support while retaining comparable verifiable content. Standard Direct Preference Optimization then distills these pairs into a single-sample policy, avoiding inference-time reranking. On patient-disjoint MIMIC-III-Ext-VeriFact-BHC test data, VERI-DPO reduces Not Supported rates from 10.7% to 1.9% under the mining verifier and from 11.6% to 6.4% under a separately prompted GPT-4o judge. In 100 blinded pairwise assessments by two domain researchers, VERI-DPO is preferred over the base model 56 times versus 18 times for factual faithfulness. In a locked zero-shot MIMIC-IV-Ext-BHC transfer test with 1,000 patients and no model adaptation, VERI-DPO lowers Not Supported rates with nearly unchanged scored-claim counts. Multi-seed ablations show that verifier-guided pair construction drives the gains, while coverage and anti-degeneration controls prevent apparent factuality improvements from coming from shorter or less checkable outputs.

Weixin Liu, Congning Ni, Qingyuan Song +4

Vanderbilt University, Nashville, TN, USA · Vanderbilt University Medical Center, Nashville, TN, USA · Virginia Tech, Blacksburg, VA, USA

Category-Aware Atomic Claims Evidence-Grounded Questions Summarization Preference Alignment Learning

Mar 4, 2026cs.AI

Towards Natural Personalization: Evaluating Long-Horizon Preference Following in Personalized User-LLM Interactions

Large Language Models (LLMs) are increasingly serving as personal assistants, where users may share individual preferences over extended interactions. However, assessing how well LLMs can follow these preferences in natural, long-term situations remains underexplored. This work proposes RealPref, a benchmark for evaluating natural preference-following in personalized user-LLM interactions. RealPref features 100 synthetic user profiles, 1300 personalized preferences, 4 types of preference expression (from explicit to implicit), and long-horizon interaction histories. It explored three types of test tasks (multiple-choice, true-or-false, and open-ended), with granular rubrics for LLM-as-a-judge evaluation. Results indicate that LLM performance drops significantly as context length grows and preference expression becomes more implicit, and that generalizing user preference understanding to unseen scenarios poses further challenges. RealPref and these findings provide a foundation for future research to develop user-aware LLM assistants that better adapt to individual needs.

Qianyun Guo, Yibo Li, Yue Liu +1

National University of Singapore, Singapore.

Large Language Model Personalization Preference Alignment Learning Personalization Self

Feb 19, 2026cs.LG

MARS: Margin and Semantic-Aware Data Augmentation for Reward Modeling

Reward modeling is central to alignment pipelines such as RLHF, RLAIF, and PPO-based policy optimization, yet its reliability is constrained by limited and heterogeneous human preference data that are expensive to collect at scale. While synthetic augmentation can expand preference supervision, existing methods often augment uniformly or at the representation level, without targeting examples where the reward model is uncertain or prone to mis-ranking. In this paper, we introduce MARS (Margin and Semantic-Aware Data Augmentation for Reward Modeling), an adaptive augmentation framework that prioritizes low-margin preference pairs and uses semantic distance as a second layer for refinement to enhance the contrast between the chosen and rejected responses. Across multiple preference datasets, reward-model backbones, downstream alignment settings, and benchmarks including RewardBench and AlpacaEval, MARS improves both reward-model quality and alignment performance over existing baselines. Our results show that reward-model augmentation is most effective when guided by both model margins and semantic structure.

Payel Bhattacharjee, Osvaldo Simeone, Ravi Tandon

University of Arizona · Tucson, AZ, USA · Northeastern University London +1

Progress Reward Modeling Augmentation Preference Alignment Learning