mlsanityData to 14 Oct 2026

Inducing Reasoning Primitives from Agent Traces

arXiv 2606.02994 ↗First posted 2 Jun 2026cs.AI cs.CLcited by 0

On the map: LLM reasoning → Reasoning for agents, planning and code

also in 1 more
Abstract (arXiv)

Existing methods induce various forms of knowledge from agent traces, from sample workflows to programmatic skills. We propose ARISE (Agent Routine Induction from Successful Experience), which identifies recurring procedures in agent traces and synthesizes them into a reusable library of agent subroutines. Across five tasks in narrative deduction, rule application, and constraint planning, it is shown to be a more effective learning strategy than prompting methods and prose and code artifact induction methods. ARISE outperforms chain-of-thought in all ten model-task settings of the main evaluation. Compared to state-of-the-art methods in agent experience induction, our method scores 6.1 percentage points over the strongest evaluated baselines, and 11.5 percentage points over code-skill induction. In our controlled experiment, ARISE also improves average accuracy by up to 15.6 percentage points compared to formatting the induced knowledge as skills and by up to 19.1 points over rewriting it as code tools.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 39 of its 42 references that are in the map. First posted on arXiv 2 Jun 2026.

Builds on

39 of 42 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024AFlow: Automating Agentic Workflow Generation14 Oct
  2. 2023Language Agent Tree Search Unifies Reasoning Acting and Planning in Language Models6 Oct

Other references · 37

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  3. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  4. 2023Reflexion: Language Agents with Verbal Reinforcement Learningcited by 4,276
  5. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  6. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  7. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Modelscited by 3,211
  8. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,191
  9. 2022Code as Policies: Language Model Programs for Embodied Controlcited by 1,210
  10. 2022Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Taskscited by 1,076
  11. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 940
  12. 2023ExpeL: LLM Agents Are Experiential Learnerscited by 876
25 more
  1. 2022PAL: Program-aided Language Modelscited by 652
  2. 2023Plan-and-Solve Prompting: Improving Zero-Shot Chain-of-Thought Reasoning by Large Language Modelscited by 560
  3. 2022Decomposed Prompting: A Modular Approach for Solving Complex Taskscited by 558
  4. 2025GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningcited by 542
  5. 2024Training Language Models to Self-Correct via Reinforcement Learningcited by 426
  6. 2024Agent Workflow Memorycited by 383
  7. 2023Large Language Models as Tool Makerscited by 303
  8. 2024Optimizing Instructions and Demonstrations for Multi-Stage Language Model Programscited by 291
  9. 2023Faithful Chain-of-Thought Reasoningcited by 288
  10. 2024V-STaR: Training Verifiers for Self-Taught Reasonerscited by 217
  11. 2023MuSR: Testing the Limits of Chain-of-thought with Multistep Soft Reasoningcited by 201
  12. 2025SkillWeaver: Web Agents can Self-Improve by Discovering and Honing Skillscited by 192
  13. 2023CRAFT: Customizing LLMs by Creating and Retrieving from Specialized Toolsetscited by 129
  14. 2024NATURAL PLAN: Benchmarking LLMs on Natural Language Planningcited by 128
  15. 2025Inducing Programmatic Skills for Agentic Taskscited by 111
  16. 2024Self-Discover: Large Language Models Self-Compose Reasoning Structurescited by 110
  17. 2024TroVE: Inducing Verifiable and Efficient Toolboxes for Solving Programmatic Taskscited by 75
  18. 2022Top-Down Synthesis for Library Learningcited by 58
  19. 2024VLM Agents Generate Their Own Memories: Distilling Experience into Embodied Programs of Thoughtcited by 55
  20. 2023LILO: Learning Interpretable Libraries by Compressing and Documenting Codecited by 49
  21. 2024RuleArena: A Benchmark for Rule-Guided Reasoning with LLMs in Real-World Scenarioscited by 27
  22. 2025RLAD: Training LLMs to Discover Abstractions for Solving Reasoning Problemscited by 21
  23. 2026Internalizing LLM Reasoning via Discovery and Replay of Latent Actionscited by 4
  24. 2024Watch Your Steps: Observable and Modular Chains of Thoughtcited by 3
  25. 2025Semi-structured LLM Reasoners Can Be Rigorously Auditedcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Second Thought: Reasoning in Parallel as LLM Agents Act and Observe
  2. 2026AgentEscapeBench: Evaluating Out-of-Domain Tool-Grounded Reasoning in LLM Agents
  3. 2024Textualized Agent-Style Reasoning for Complex Tasks by Multiple Round LLM Generation
  4. 2026Thinking Before Thinking: Scaling Agentic Inference Through Meta-Reasoning
  5. 2024On the Brittle Foundations of ReAct Prompting for Agentic Large Language Models
  6. 2026Project Ariadne: A Structural Causal Framework for Auditing Faithfulness in LLM Agents
  7. 2025CodeARC: Benchmarking Reasoning Capabilities of LLM Agents for Inductive Program Synthesis
  8. 2025Reasoning on a Budget: A Survey of Adaptive and Controllable Test-Time Compute in LLMs