Iterative Deployment Improves Planning Skills in LLMs

• URL: http://arxiv.org/abs/2512.24940v1
• Date: Wed, 31 Dec 2025 16:03:14 GMT
• Title: Iterative Deployment Improves Planning Skills in LLMs
• Abstract summary: We show that iterative deployment of large language models (LLMs) can significantly change the properties of the resultant models.<n>By testing this mechanism on various planning domains, we observe substantial improvements in planning skills.<n>We then provide theoretical analysis showing that iterative deployment effectively implements reinforcement learning (RL) training in the outer-loop.
• Score: 36.668436209940516
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We show that iterative deployment of large language models (LLMs), each fine-tuned on data carefully curated by users from the previous models' deployment, can significantly change the properties of the resultant models. By testing this mechanism on various planning domains, we observe substantial improvements in planning skills, with later models displaying emergent generalization by discovering much longer plans than the initial models. We then provide theoretical analysis showing that iterative deployment effectively implements reinforcement learning (RL) training in the outer-loop (i.e. not as part of intentional model training), with an implicit reward function. The connection to RL has two important implications: first, for the field of AI safety, as the reward function entailed by repeated deployment is not defined explicitly, and could have unexpected implications to the properties of future model deployments. Second, the mechanism highlighted here can be viewed as an alternative training regime to explicit RL, relying on data curation rather than explicit rewards.

Related papers

• Large Vision-Language Models Get Lost in Attention [51.851592109135716]
  We propose a unified framework grounded in information theory and geometry to quantify the geometric and entropic nature of residual updates.<n>Applying this unified framework reveals a fundamental functional decoupling: Attention acts as a subspace-language operator focused on reconfiguration, whereas FFNs serve as subspace-language operators driving semantic innovation.
  arXiv  Detail & Related papers  (2026-05-07T04:45:52Z)
• Why Does Reinforcement Learning Generalize? A Feature-Level Mechanistic Study of Post-Training in Large Language Models [51.45122910254346]
  Reinforcement learning (RL)-based post-training often improves the reasoning performance of large language models beyond the training domain.<n>Supervised fine-tuning (SFT) frequently leads to general forgetting capabilities.<n>We present a feature-level mechanistic analysis methodology to probe RL generalization using a controlled experimental setup.
  arXiv  Detail & Related papers  (2026-04-27T21:22:34Z)
• Regularized Latent Dynamics Prediction is a Strong Baseline For Behavioral Foundation Models [35.088440282359024]
  Behavioral Foundation Models (BFMs) produce agents with the capability to adapt to any unknown reward or task.<n>These methods are only able to produce near-optimal policies for the reward functions that are in the span of some pre-existing state features.<n>We propose an approach, Regularized Latent Dynamics Prediction (RLDP), that can match or surpass state-of-the-art complex representation learning methods for zero-shot RL.
  arXiv  Detail & Related papers  (2026-03-16T19:39:27Z)
• Generative Actor Critic [74.04971271003869]
  Generative Actor Critic (GAC) is a novel framework that decouples sequential decision-making by reframing textitpolicy evaluation as learning a generative model of the joint distribution over trajectories and returns.<n>Experiments on Gym-MuJoCo and Maze2D benchmarks demonstrate GAC's strong offline performance and significantly enhanced offline-to-online improvement compared to state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-12-25T06:31:11Z)
• Spatiotemporal Forecasting as Planning: A Model-Based Reinforcement Learning Approach with Generative World Models [45.523937630646394]
  We propose SFP Forecasting as Planning (SFP), a new paradigm in Model Based Reinforcement Learning.<n>SFP constructs a novel World Model to simulate diverse high-temporal future states, enabling an "imagination-based" environmental simulation.
  arXiv  Detail & Related papers  (2025-10-05T03:57:38Z)
• SHAP-Guided Regularization in Machine Learning Models [1.0515439489916734]
  We propose a SHAP-guided regularization framework that incorporates feature importance constraints into model training to enhance both predictive performance and interpretability.<n>Our approach applies entropy-based penalties to encourage sparse, concentrated feature attributions while promoting stability across samples.
  arXiv  Detail & Related papers  (2025-07-31T15:45:38Z)
• Behavior Injection: Preparing Language Models for Reinforcement Learning [45.744838898763554]
  We analyze the per-step influence of the RL objective and identify two key conditions for effective post-training.<n>We propose behavior injection, a task-agnostic data augmentation scheme applied prior to RL.<n>We evaluate our method across two reasoning benchmarks with multiple base models.
  arXiv  Detail & Related papers  (2025-05-25T00:54:50Z)
• LARES: Latent Reasoning for Sequential Recommendation [96.26996622771593]
  We present LARES, a novel and scalable LAtent REasoning framework for Sequential recommendation.<n>Our proposed approach employs a recurrent architecture that allows flexible expansion of reasoning depth without increasing parameter complexity.<n>Our framework exhibits seamless compatibility with existing advanced models, further improving their recommendation performance.
  arXiv  Detail & Related papers  (2025-05-22T16:22:54Z)
• Echo Chamber: RL Post-training Amplifies Behaviors Learned in Pretraining [74.83412846804977]
  Reinforcement learning (RL)-based fine-tuning has become a crucial step in post-training language models.<n>We present a systematic end-to-end study of RL fine-tuning for mathematical reasoning by training models entirely from scratch.
  arXiv  Detail & Related papers  (2025-04-10T17:15:53Z)
• On the Modeling Capabilities of Large Language Models for Sequential Decision Making [52.128546842746246]
  Large pretrained models are showing increasingly better performance in reasoning and planning tasks. We evaluate their ability to produce decision-making policies, either directly, by generating actions, or indirectly. In environments with unfamiliar dynamics, we explore how fine-tuning LLMs with synthetic data can significantly improve their reward modeling capabilities.
  arXiv  Detail & Related papers  (2024-10-08T03:12:57Z)
• INFOrmation Prioritization through EmPOWERment in Visual Model-Based RL [90.06845886194235]
  We propose a modified objective for model-based reinforcement learning (RL) We integrate a term inspired by variational empowerment into a state-space model based on mutual information. We evaluate the approach on a suite of vision-based robot control tasks with natural video backgrounds.
  arXiv  Detail & Related papers  (2022-04-18T23:09:23Z)
• Forethought and Hindsight in Credit Assignment [62.05690959741223]
  We work to understand the gains and peculiarities of planning employed as forethought via forward models or as hindsight operating with backward models. We investigate the best use of models in planning, primarily focusing on the selection of states in which predictions should be (re)-evaluated.
  arXiv  Detail & Related papers  (2020-10-26T16:00:47Z)

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.