Recon: Reconstruction-Guided Reasoning Synthesis for User Modeling

• URL: http://arxiv.org/abs/2605.26969v1
• Date: Tue, 26 May 2026 12:55:11 GMT
• Title: Recon: Reconstruction-Guided Reasoning Synthesis for User Modeling
• Abstract summary: We propose Recon, which uses action reconstruction to score reasoning traces by their predictive power.<n>We show that Recon-synthesized reasoning transfers across models, and improves user modeling beyond the reconstruction model.
• Score: 30.893319122518587
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: User modeling aims to use language models (LMs) to mimic an individual's behavior from a corpus of past context-action pairs (e.g., conversation turns), enabling the simulation of users in settings like behavioral science, human-AI collaboration, and market research. Recent approaches augment these corpora with synthesized reasoning traces, typically generated by conditioning on both context and action. However, such conditioning constitutes post-hoc rationalization rather than reasoning: the trace is guaranteed to justify the action, but may not encode the underlying latent causal decision paths. We propose Recon, which uses action reconstruction to score reasoning traces by their predictive power: given a context and candidate reasoning, a reconstruction model predicts the action, and reconstruction fidelity determines reasoning quality. Across four domains, Recon achieves a 54.7% win rate over Backward Synthesis, a standard post-hoc rationalization baseline. Further, we find that training a reasoning synthesis model with rewards derived from Recon improves downstream user modeling performance, achieving a win rate of up to 70.0% over baselines. We further show that Recon-synthesized reasoning transfers across models, and improves user modeling beyond the reconstruction model. Our work demonstrates that post-hoc rationalization is insufficient for reasoning synthesis, and that useful and interpretable reasoning should naturally elicit the action from the context.

Related papers

• RationalRewards: Reasoning Rewards Scale Visual Generation Both Training and Test Time [51.256131853751754]
  We show that teaching reward models to produce explicit, multi-dimensional critiques transforms them from passive evaluators into active optimization tools.<n>We introduce Preference-Anchored Rationalization (PARROT), a principled framework that recovers high-quality rationales from readily available preference data.<n>The resulting model, RationalRewards (8B), achieves state-of-the-art preference prediction among open-source reward models.
  arXiv  Detail & Related papers  (2026-04-13T15:38:09Z)
• Structured Reasoning for Large Language Models [59.215789462977206]
  We propose Structured Reasoning (SCR), a framework that decouples reasoning trajectories into explicit, evaluable, and trainable components.<n>SCR substantially improves reasoning efficiency and self-verification.<n>Compared with existing reasoning paradigms, it reduces output token length by up to 50%.
  arXiv  Detail & Related papers  (2026-01-12T04:04:01Z)
• ROVER: Benchmarking Reciprocal Cross-Modal Reasoning for Omnimodal Generation [79.17352367219736]
  ROVER tests the use of one modality to guide, verify, or refine outputs in the other.<n>ROVER is a human-annotated benchmark that explicitly targets reciprocal cross-modal reasoning.
  arXiv  Detail & Related papers  (2025-11-03T02:27:46Z)
• OpenRubrics: Towards Scalable Synthetic Rubric Generation for Reward Modeling and LLM Alignment [38.1645520104553]
  We introduce OpenRubrics, a large-scale collection of (prompt,explicit) pairs for training rubric-generation and rubric-based reward models.<n>To elicit discriminative and comprehensive evaluation signals, we introduce Contrastive Generation (CRG), which derives both hard rules (implicit qualities) and principles (implicit qualities) by contrasting preferred and rejected responses.<n>Our results show that rubrics provide scalable alignment signals that narrow the gap between costly human evaluation and automated reward modeling.
  arXiv  Detail & Related papers  (2025-10-09T03:31:26Z)
• Controlled Agentic Planning & Reasoning for Mechanism Synthesis [18.8323743697237]
  This work presents a dual-agent acllm-based reasoning framework for automated planar mechanism synthesis.<n>From a natural-language task description, the system composes symbolic constraints and equations, generates and parametrises simulation code, and iteratively refines designs via critic-driven feedback.
  arXiv  Detail & Related papers  (2025-05-23T08:16:32Z)
• RM-R1: Reward Modeling as Reasoning [81.50471199906738]
  Reasoning Reward Models (ReasRMs) formulate reward modeling as a reasoning task.<n>We propose a reasoning-oriented training pipeline and train a family of ReasRMs, RM-R1.<n>Our models achieve state-of-the-art performance across three reward model benchmarks on average.
  arXiv  Detail & Related papers  (2025-05-05T06:11:12Z)
• Model Reconstruction Using Counterfactual Explanations: A Perspective From Polytope Theory [9.771997770574947]
  We analyze how model reconstruction using counterfactuals can be improved. Our main contribution is to derive novel theoretical relationships between the error in model reconstruction and the number of counterfactual queries.
  arXiv  Detail & Related papers  (2024-05-08T18:52:47Z)
• RewardBench: Evaluating Reward Models for Language Modeling [100.28366840977966]
  We present RewardBench, a benchmark dataset and code-base for evaluation of reward models. The dataset is a collection of prompt-chosen-rejected trios spanning chat, reasoning, and safety. On the RewardBench leaderboard, we evaluate reward models trained with a variety of methods.
  arXiv  Detail & Related papers  (2024-03-20T17:49:54Z)
• SocREval: Large Language Models with the Socratic Method for Reference-Free Reasoning Evaluation [78.23119125463964]
  We develop SocREval, a novel approach for prompt design in reference-free reasoning evaluation.<n>SocREval significantly improves GPT-4's performance, surpassing existing reference-free and reference-based reasoning evaluation metrics.
  arXiv  Detail & Related papers  (2023-09-29T18:25:46Z)
• Rethinking Coherence Modeling: Synthetic vs. Downstream Tasks [15.044192886215887]
  Coherence models are typically evaluated only on synthetic tasks, which may not be representative of their performance in downstream applications. We conduct experiments on benchmarking well-known traditional and neural coherence models on synthetic sentence ordering tasks. Our results demonstrate a weak correlation between the model performances in the synthetic tasks and the downstream applications.
  arXiv  Detail & Related papers  (2020-04-30T08:00:42Z)

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.