ARCANE: A Multi-Agent Framework for Interpretable and Configurable Alignment

• URL: http://arxiv.org/abs/2512.06196v1
• Date: Fri, 05 Dec 2025 22:39:54 GMT
• Title: ARCANE: A Multi-Agent Framework for Interpretable and Configurable Alignment
• Abstract summary: ARCANE is a framework that frames alignment as a multi-agent collaboration problem that dynamically represents stakeholder preferences as natural-language rubrics.<n>Inspired by utility theory, we formulate rubric learning as a reconstruction problem and apply a regularized Group-Sequence Policy Optimization (GSPO) procedure.<n>Our results show that rubric-based reward models offer a promising path toward interpretable, test-time adaptive alignment for complex, long-horizon AI systems.
• Score: 11.754819989978785
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As agents based on large language models are increasingly deployed to long-horizon tasks, maintaining their alignment with stakeholder preferences becomes critical. Effective alignment in such settings requires reward models that are interpretable so that stakeholders can understand and audit model objectives. Moreover, reward models must be capable of steering agents at interaction time, allowing preference shifts to be incorporated without retraining. We introduce ARCANE, a framework that frames alignment as a multi-agent collaboration problem that dynamically represents stakeholder preferences as natural-language rubrics: weighted sets of verifiable criteria that can be generated on-the-fly from task context. Inspired by utility theory, we formulate rubric learning as a reconstruction problem and apply a regularized Group-Sequence Policy Optimization (GSPO) procedure that balances interpretability, faithfulness, and computational efficiency. Using a corpus of 219 labeled rubrics derived from the GDPVal benchmark, we evaluate ARCANE on challenging tasks requiring multi-step reasoning and tool use. The learned rubrics produce compact, legible evaluations and enable configurable trade-offs (e.g., correctness vs. conciseness) without retraining. Our results show that rubric-based reward models offer a promising path toward interpretable, test-time adaptive alignment for complex, long-horizon AI systems.

Related papers

• From Profiling to Synthesis: Benchmarking Implicit Behavioral Alignment in Personalized LLM Agents [56.187590188698955]
  Large Language Models have enabled increasingly capable autonomous agents, yet personalization remains critical for making such agents practically useful.<n>Recent benchmarks have begun evaluating personalization in agents, but they largely rely on static preference snapshots, fixed interaction logs, or question answering over predefined user profiles.<n>We introduce IBA-Bench, a benchmark for implicit behavioral alignment constructed from longitudinal interaction histories.<n>We propose IBA-Agent, an agent framework that reconciles conflicting priorities through broad retrieval and trajectory-level alignment.
  arXiv  Detail & Related papers  (2026-08-03T12:52:04Z)
• REAR: Test-time Preference Realignment through Reward Decomposition [68.09214603569744]
  We introduce a novel framework that models the task as a realignment problem, since the base model often fails to align with the stated preference.<n>We derive a REAlignment Reward (REAR) that selectively rescales the proportions of these two reward terms.<n>We show that REAR can be formulated as a linear combination of token-level policy log-probabilities, making it computationally efficient and easy to integrate with various TTS algorithms.
  arXiv  Detail & Related papers  (2026-06-29T14:17:53Z)
• The Art of Interrogation: Consistency Amplifies Factuality in Spatial Reasoning [65.67266333751569]
  Current Large Reasoning Models (LRMs) exhibit remarkable general capabilities but significantly underperform in spatial reasoning tasks.<n>We propose a self-supervised reinforcement learning framework that targets the internal reasoning process without requiring ground-truth annotations.<n>We show that this label-free consistency training approaches the accuracy of models trained with ground-truth supervision and achieves similar generalization across diverse tasks and data domains.
  arXiv  Detail & Related papers  (2026-06-10T10:50:06Z)
• TRACE: A Unified Rollout Budget Allocation Framework for Efficient Agentic Reinforcement Learning [52.67756371231985]
  Reinforcement learning with verifiable rewards (RLVR) is a promising approach for enhancing reasoning and agentic behavior in large language models.<n>We introduce Tree Rollout Allocation for Contrastive Exploration (TRACE), a unified rollout allocation framework.<n>Technically, TRACE allocates rollout budget to both prompt roots and intermediate prefixes that are most likely to yield mixed terminal rewards.
  arXiv  Detail & Related papers  (2026-06-09T17:16:03Z)
• Steered Generation via Gradient-Based Optimization on Sparse Query Features [1.587869707099313]
  Prototype-Based Sparse Steering is a framework that applies Sparse Autoencoders specifically to query activations.<n>We demonstrate that sparse query representations provide the necessary disentanglement for unified, interpretable control over both logical planning and stylistic nuance.
  arXiv  Detail & Related papers  (2026-05-21T21:13:14Z)
• Co-evolving Agent Architectures and Interpretable Reasoning for Automated Optimization [6.428928591765431]
  Operations research with large language models (LLMs) remains limited by hand-crafted reasoning.<n>We propose EvoOR-Agent, a co-evolutionary framework for automated optimization.<n>Case studies and ablation analyses indicate that explicit architecture evolution and graph-supported reasoning-trajectory search contribute to both performance improvement and structural interpretability.
  arXiv  Detail & Related papers  (2026-04-20T01:44:18Z)
• RPA-Check: A Multi-Stage Automated Framework for Evaluating Dynamic LLM-based Role-Playing Agents [4.07447364754644]
  This paper introduces RPA-Check, a multi-stage automated evaluation framework designed to objectively assess the performance of LLM-based RPAs.<n>We validate this framework by applying it to LLM Court, a serious game for forensic training involving several quantized local models.
  arXiv  Detail & Related papers  (2026-04-13T16:08:03Z)
• Cluster-R1: Large Reasoning Models Are Instruction-following Clustering Agents [81.79110139097297]
  General-purpose embedding models excel at recognizing semantic similarities but fail to capture the characteristics of texts specified by user instructions.<n>We reframe instruction-following clustering as a generative task and train large reasoning models as autonomous clustering agents.<n>Our reasoning-driven training pipeline enables LRMs to interpret high-level clustering instructions and then infer the corresponding latent groupings.
  arXiv  Detail & Related papers  (2026-03-06T10:15:54Z)
• Optimizing In-Context Demonstrations for LLM-based Automated Grading [31.353360036776976]
  GUIDE (Grading Using Iteratively Designed Exemplars) is a framework that reframes exemplar selection and refinement as a boundary-focused optimization problem.<n>We show that GUIDE significantly outperforms standard retrieval baselines in experiments in physics, chemistry, and pedagogical content knowledge.
  arXiv  Detail & Related papers  (2026-02-28T04:52:38Z)
• Towards Adaptive, Scalable, and Robust Coordination of LLM Agents: A Dynamic Ad-Hoc Networking Perspective [31.81236449944822]
  RAPS is a reputation-aware publish-subscribe paradigm for adaptive, scalable, and robust coordination of LLM agents.<n>RAPS incorporates two coherent overlays: (i) Reactive Subscription, enabling agents to dynamically refine their intents; and (ii) Bayesian Reputation, empowering each agent with a local watchdog to detect and isolate malicious peers.
  arXiv  Detail & Related papers  (2026-02-08T15:26:02Z)
• MAESTRO: Meta-learning Adaptive Estimation of Scalarization Trade-offs for Reward Optimization [56.074760766965085]
  Group-Relative Policy Optimization has emerged as an efficient paradigm for aligning Large Language Models (LLMs)<n>We propose MAESTRO, which treats reward scalarization as a dynamic latent policy, leveraging the model's terminal hidden states as a semantic bottleneck.<n>We formulate this as a contextual bandit problem within a bi-level optimization framework, where a lightweight Conductor network co-evolves with the policy by utilizing group-relative advantages as a meta-reward signal.
  arXiv  Detail & Related papers  (2026-01-12T05:02:48Z)
• CoT Referring: Improving Referring Expression Tasks with Grounded Reasoning [67.18702329644526]
  CoT Referring enhances model reasoning across modalities through a structured, chain-of-thought training data structure.<n>We restructure the training data to enforce a new output form, providing new annotations for existing datasets.<n>We also integrate detection and segmentation capabilities into a unified MLLM framework, training it with a novel adaptive weighted loss to optimize performance.
  arXiv  Detail & Related papers  (2025-10-03T08:50:21Z)
• Visual Document Understanding and Question Answering: A Multi-Agent Collaboration Framework with Test-Time Scaling [83.78874399606379]
  We propose MACT, a Multi-Agent Collaboration framework with Test-Time scaling.<n>It comprises four distinct small-scale agents, with clearly defined roles and effective collaboration.<n>It shows superior performance with a smaller parameter scale without sacrificing the ability of general and mathematical tasks.
  arXiv  Detail & Related papers  (2025-08-05T12:52:09Z)
• Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation [91.17994756436259]
  Multi-agent systems (MAS) based on large language models (LLMs) have emerged as a powerful solution for dealing with complex problems across diverse domains.<n>Existing approaches are fundamentally constrained by their reliance on a template graph modification paradigm with a predefined set of agents and hard-coded interaction structures.<n>We propose ARG-Designer, a novel autoregressive model that operationalizes this paradigm by constructing the collaboration graph from scratch.
  arXiv  Detail & Related papers  (2025-07-24T09:17:41Z)
• Parallelism Meets Adaptiveness: Scalable Documents Understanding in Multi-Agent LLM Systems [0.8437187555622164]
  Large language model (LLM) agents have shown increasing promise for collaborative task completion.<n>Existing multi-agent frameworks often rely on static, fixed roles, and limited inter-agent communication.<n>This paper proposes a coordination framework that enables adaptiveness through three core mechanisms.
  arXiv  Detail & Related papers  (2025-07-22T22:42:51Z)
• SCoRE: Streamlined Corpus-based Relation Extraction using Multi-Label Contrastive Learning and Bayesian kNN [0.2812395851874055]
  We introduce SCoRE, a modular and cost-effective sentence-level relation extraction system.<n>SCoRE enables easy PLM switching, requires no finetuning, and adapts smoothly to diverse corpora and KGs.<n>We show that SCoRE matches or surpasses state-of-the-art methods while significantly reducing energy consumption.
  arXiv  Detail & Related papers  (2025-07-09T14:33:07Z)
• Tree-Based Text Retrieval via Hierarchical Clustering in RAGFrameworks: Application on Taiwanese Regulations [0.0]
  We propose a hierarchical clustering-based retrieval method that eliminates the need to predefine k.<n>Our approach maintains the accuracy and relevance of system responses while adaptively selecting semantically relevant content.<n>Our framework is simple to implement and easily integrates with existing RAG pipelines, making it a practical solution for real-world applications under limited resources.
  arXiv  Detail & Related papers  (2025-06-16T15:34:29Z)
• MA-RAG: Multi-Agent Retrieval-Augmented Generation via Collaborative Chain-of-Thought Reasoning [36.3918410061572]
  MA-RAG addresses the inherent ambiguities and reasoning challenges in complex information-seeking tasks.<n>Unlike conventional RAG methods that rely on end-to-end fine-tuning or isolated component enhancements, MA-RAG orchestrates a collaborative set of specialized AI agents.<n>Our results highlight the effectiveness of collaborative, modular reasoning in retrieval-augmented systems.
  arXiv  Detail & Related papers  (2025-05-26T15:05:18Z)
• Dynamic benchmarking framework for LLM-based conversational data capture [0.0]
  This paper introduces a benchmarking framework to assess large language models (LLMs)<n>It integrates generative agent simulation to evaluate performance on key dimensions: information extraction, context awareness, and adaptive engagement.<n>Results show that adaptive strategies improve data extraction accuracy, especially when handling ambiguous responses.
  arXiv  Detail & Related papers  (2025-02-04T15:47:47Z)
• CELA: Cost-Efficient Language Model Alignment for CTR Prediction [70.65910069412944]
  Click-Through Rate (CTR) prediction holds a paramount position in recommender systems.<n>Recent efforts have sought to mitigate these challenges by integrating Pre-trained Language Models (PLMs)<n>We propose textbfCost-textbfEfficient textbfLanguage Model textbfAlignment (textbfCELA) for CTR prediction.
  arXiv  Detail & Related papers  (2024-05-17T07:43:25Z)
• Entropy-Regularized Token-Level Policy Optimization for Language Agent Reinforcement [67.1393112206885]
  Large Language Models (LLMs) have shown promise as intelligent agents in interactive decision-making tasks. We introduce Entropy-Regularized Token-level Policy Optimization (ETPO), an entropy-augmented RL method tailored for optimizing LLMs at the token level. We assess the effectiveness of ETPO within a simulated environment that models data science code generation as a series of multi-step interactive tasks.
  arXiv  Detail & Related papers  (2024-02-09T07:45:26Z)

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.