One Model for All: Multi-Objective Controllable Language Models

• URL: http://arxiv.org/abs/2604.04497v1
• Date: Mon, 06 Apr 2026 07:48:32 GMT
• Title: One Model for All: Multi-Objective Controllable Language Models
• Abstract summary: We introduce Multi-Objective Control (MOC) to train a single language model as a preference-conditioned policy network.<n>Our approach introduces multi-objective optimization (MOO) principles into RLHF to train an LLM as a preference-conditioned policy network.<n>Extensive experiments demonstrate the advantages of MOC over baselines in three aspects.
• Score: 65.4626816393381
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Aligning large language models (LLMs) with human preferences is critical for enhancing LLMs' safety, helpfulness, humor, faithfulness, etc. Current reinforcement learning from human feedback (RLHF) mainly focuses on a fixed reward learned from average human ratings, which may weaken the adaptability and controllability of varying preferences. However, creating personalized LLMs requires aligning LLMs with individual human preferences, which is non-trivial due to the scarce data per user and the diversity of user preferences in multi-objective trade-offs, varying from emphasizing empathy in certain contexts to demanding efficiency and precision in others. Can we train one LLM to produce personalized outputs across different user preferences on the Pareto front? In this paper, we introduce Multi-Objective Control (MOC), which trains a single LLM to directly generate responses in the preference-defined regions of the Pareto front. Our approach introduces multi-objective optimization (MOO) principles into RLHF to train an LLM as a preference-conditioned policy network. We improve the computational efficiency of MOC by applying MOO at the policy level, enabling us to fine-tune a 7B-parameter model on a single A6000 GPU. Extensive experiments demonstrate the advantages of MOC over baselines in three aspects: (i) controllability of LLM outputs w.r.t. user preferences on the trade-off among multiple rewards; (ii) quality and diversity of LLM outputs, measured by the hyper-volume of multiple solutions achieved; and (iii) generalization to unseen preferences. These results highlight MOC's potential for real-world applications requiring scalable and customizable LLMs.

Related papers

• MATO: Multi-objective Personalized Alignment with Test-time Optimization for Large Language Models [47.91243135663066]
  We introduce MATO, a training-free framework for Multi-objective personalized alignment with Test-time Optimization.<n> MATO formulates personalization as a test-time optimization problem that steers the relative importance of multiple objectives through controllable weights during decoding.
  arXiv  Detail & Related papers  (2026-05-25T01:57:22Z)
• Many Preferences, Few Policies: Towards Scalable Language Model Personalization [26.263947748558824]
  holy grail of LLM personalization is a single LLM for each user, perfectly aligned with that user's preferences.<n>We develop a principled method for selecting a small portfolio of LLMs that captures representative behaviors across heterogeneous users.<n>We provide empirical results that validate these guarantees and demonstrate greater output diversity over common baselines.
  arXiv  Detail & Related papers  (2026-04-05T15:12:07Z)
• Can a Unimodal Language Agent Provide Preferences to Tune a Multimodal Vision-Language Model? [8.976163131623773]
  We propose a method that enables a language agent to give feedback to a vision-language model (VLM) to adapt text generation to the agent's preferences.<n>Using our proposed method, we find that the VLM can generate multimodal scene descriptions to help the LLM better understand multimodal context.
  arXiv  Detail & Related papers  (2026-01-10T04:28:53Z)
• Uncertainty-Aware Answer Selection for Improved Reasoning in Multi-LLM Systems [55.6590601898194]
  Large Language Models (LLMs) have demonstrated exceptional capabilities, yet selecting the most reliable response from multiple LLMs remains a challenge.<n>Existing approaches often depend on costly external verifiers, human evaluators, or self-consistency techniques that require multiple samples from a single model.<n>We propose a principled, novel and computationally efficient method to select the best response from multiple different LLMs using a calibrated log-likelihood score.
  arXiv  Detail & Related papers  (2025-09-30T01:25:19Z)
• Measuring What Makes You Unique: Difference-Aware User Modeling for Enhancing LLM Personalization [68.79814761867314]
  We propose Difference-aware Personalization Learning (DPL) to enhance Large Language Models (LLMs) personalization.<n>DPL strategically selects representative users for comparison and establishes a structured standard to extract task-relevant differences.<n>Experiments on real-world datasets demonstrate that DPL significantly enhances LLM personalization.
  arXiv  Detail & Related papers  (2025-03-04T09:53:26Z)
• LLM Bandit: Cost-Efficient LLM Generation via Preference-Conditioned Dynamic Routing [3.090041654375235]
  We present a novel framework that formulates the LLM selection process as a multi-armed bandit problem.<n>Our approach incorporates a preference-conditioned dynamic routing mechanism, allowing users to specify their preferences at inference time.<n>Our method achieves significant improvements in both accuracy and cost-effectiveness across various LLM platforms.
  arXiv  Detail & Related papers  (2025-02-04T22:09:43Z)
• Few-shot Steerable Alignment: Adapting Rewards and LLM Policies with Neural Processes [50.544186914115045]
  Large language models (LLMs) are increasingly embedded in everyday applications.<n> Ensuring their alignment with the diverse preferences of individual users has become a critical challenge.<n>We present a novel framework for few-shot steerable alignment.
  arXiv  Detail & Related papers  (2024-12-18T16:14:59Z)
• MetaAlign: Align Large Language Models with Diverse Preferences during Inference Time [50.41806216615488]
  Large Language Models (LLMs) acquire extensive knowledge and remarkable abilities from extensive text corpora. To make LLMs more usable, aligning them with human preferences is essential. We propose an effective method, textbf MetaAlign, which aims to help LLMs dynamically align with various explicit or implicit preferences specified at inference time.
  arXiv  Detail & Related papers  (2024-10-18T05:31:13Z)
• Aligning LLMs with Individual Preferences via Interaction [51.72200436159636]
  We train large language models (LLMs) that can ''interact to align''<n>We develop a multi-turn preference dataset containing 3K+ multi-turn conversations in tree structures.<n>For evaluation, we establish the ALOE benchmark, consisting of 100 carefully selected examples and well-designed metrics to measure the customized alignment performance during conversations.
  arXiv  Detail & Related papers  (2024-10-04T17:48:29Z)
• Beyond One-Preference-Fits-All Alignment: Multi-Objective Direct Preference Optimization [76.09576643028362]
  We present Multi-Objective Direct Preference Optimization (MODPO) for multiple alignment objectives. MODPO folds language modeling directly into reward modeling, training language models as implicit collective reward models. It theoretically yields the same optimal solutions as MORLHF but is practically more stable and efficient.
  arXiv  Detail & Related papers  (2023-10-05T17:35:26Z)

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

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