Expanding the Action Space of LLMs to Reason Beyond Language

• URL: http://arxiv.org/abs/2510.07581v2
• Date: Fri, 17 Oct 2025 13:22:31 GMT
• Title: Expanding the Action Space of LLMs to Reason Beyond Language
• Abstract summary: Large Language Models (LLMs) are powerful reasoners in natural language.<n>Their actions are typically confined to outputting vocabulary tokens.<n>This overloads the model's language with both reasoning and control duties.
• Score: 17.24378705018116
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Models (LLMs) are powerful reasoners in natural language, but their actions are typically confined to outputting vocabulary tokens. As a result, interactions with external environments -- such as symbolic operators or simulators -- must be expressed through text in predefined formats, parsed, and routed to external interfaces. This overloads the model's language with both reasoning and control duties, and requires a hand-crafted parser, external to the LLM. To address this, we decouple environment interactions from language by internalizing them in an Expanded Action space (ExpA), beyond the vocabulary. The model starts reasoning in the default language environment, but may trigger routing actions and switch to an external environment at any time. From there, the model can only invoke environment-specific actions, receive feedback from the environment, and potentially route back to language as a result. To promote effective exploration of the expanded action space and new environments, we introduce ExpA Reinforcement Learning (EARL) with counterfactual policy optimization. On tasks requiring multi-turn interactions and contingent planning, EARL outperforms strong baselines with vocabulary-constrained actions. It performs robustly across calculator-based multi-task learning and, in the partially observed sorting problem, achieves perfect Sort-4 accuracy while self-discovering an efficient algorithm competitive with classical designs.

Related papers

• Behavior Tokens Speak Louder: Disentangled Explainable Recommendation with Behavior Vocabulary [22.925582428795437]
  BEAT is a framework that tokenizes user and item behaviors into discrete, interpretable sequences.<n>We show that BEAT improves zero-shot recommendation performance while generating coherent and informative explanations.
  arXiv  Detail & Related papers  (2025-12-17T17:24:24Z)
• Sample-Efficient Online Learning in LM Agents via Hindsight Trajectory Rewriting [92.57796055887995]
  We introduce ECHO, a prompting framework that adapts hindsight experience replay from reinforcement learning for language model agents.<n> ECHO generates optimized trajectories for alternative goals that could have been achieved during failed attempts.<n>We evaluate ECHO on stateful versions of XMiniGrid, a text-based navigation and planning benchmark, and PeopleJoinQA, a collaborative information-gathering enterprise simulation.
  arXiv  Detail & Related papers  (2025-10-11T18:11:09Z)
• From What to Why: Thought-Space Recommendation with Small Language Models [2.134948383299948]
  Large Language Models (LLMs) have advanced recommendation capabilities through enhanced reasoning, but pose significant challenges for real-world deployment.<n>Small Language Models (SLMs) offer an efficient alternative, their reasoning capabilities for recommendation remain underexplored.<n>We propose PULSE (Preference Understanding by Latent Semantic Embeddings), a framework that treats SLM-generated rationales as director learning signals.
  arXiv  Detail & Related papers  (2025-10-08T11:22:26Z)
• CodeDiffuser: Attention-Enhanced Diffusion Policy via VLM-Generated Code for Instruction Ambiguity [23.77040677368575]
  We introduce a novel robotic manipulation framework that can accomplish tasks specified by potentially ambiguous natural language.<n>This framework employs a Vision-Language Model (VLM) to interpret abstract concepts in natural language instructions.<n>We show that our approach excels across challenging manipulation tasks involving language ambiguity, contact-rich manipulation, and multi-object interactions.
  arXiv  Detail & Related papers  (2025-06-19T23:42:03Z)
• Plant in Cupboard, Orange on Rably, Inat Aphone. Benchmarking Incremental Learning of Situation and Language Model using a Text-Simulated Situated Environment [18.256529559741075]
  Large Language Models (LLMs) serve as key components in agent systems, where their common-sense knowledge significantly impacts performance as language-based planners for situated or embodied action.<n>We assess LLMs' incremental learning (based on feedback from the environment), and controlled in-context learning abilities using a text-based environment.<n>Results show that larger commercial models have a substantial gap in performance compared to open-weight but almost all models struggle with the synthetic words experiments.
  arXiv  Detail & Related papers  (2025-02-17T12:20:39Z)
• Language Agents Meet Causality -- Bridging LLMs and Causal World Models [50.79984529172807]
  We propose a framework that integrates causal representation learning with large language models. This framework learns a causal world model, with causal variables linked to natural language expressions. We evaluate the framework on causal inference and planning tasks across temporal scales and environmental complexities.
  arXiv  Detail & Related papers  (2024-10-25T18:36:37Z)
• From LLMs to Actions: Latent Codes as Bridges in Hierarchical Robot Control [58.72492647570062]
  We introduce our method -- Learnable Latent Codes as Bridges (LCB) -- as an alternate architecture to overcome limitations.<n>We find that methodoutperforms baselines that leverage pure language as the interface layer on tasks that require reasoning and multi-step behaviors.
  arXiv  Detail & Related papers  (2024-05-08T04:14:06Z)
• Can large language models explore in-context? [87.49311128190143]
  We deploy Large Language Models as agents in simple multi-armed bandit environments. We find that the models do not robustly engage in exploration without substantial interventions.
  arXiv  Detail & Related papers  (2024-03-22T17:50:43Z)
• Integrating LLMs and Decision Transformers for Language Grounded Generative Quality-Diversity [0.0]
  Quality-Diversity is a branch of optimization that is often applied to problems from the Reinforcement Learning and control domains. We propose a Large Language Model to augment the repertoire with natural language descriptions of trajectories. We also propose an LLM-based approach to evaluating the performance of such generative agents.
  arXiv  Detail & Related papers  (2023-08-25T10:00:06Z)
• Augmented Language Models: a Survey [55.965967655575454]
  This survey reviews works in which language models (LMs) are augmented with reasoning skills and the ability to use tools. We refer to them as Augmented Language Models (ALMs) The missing token objective allows ALMs to learn to reason, use tools, and even act, while still performing standard natural language tasks.
  arXiv  Detail & Related papers  (2023-02-15T18:25:52Z)
• Language Models as Zero-Shot Planners: Extracting Actionable Knowledge for Embodied Agents [111.33545170562337]
  We investigate the possibility of grounding high-level tasks, expressed in natural language, to a chosen set of actionable steps. We find that if pre-trained LMs are large enough and prompted appropriately, they can effectively decompose high-level tasks into low-level plans. We propose a procedure that conditions on existing demonstrations and semantically translates the plans to admissible actions.
  arXiv  Detail & Related papers  (2022-01-18T18:59:45Z)

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.