mlsanityData to 14 Oct 2026

Separable Pathways for Causal Reasoning: How Architectural Scaffolding Enables Hypothesis-Space Restructuring in LLM Agents

arXiv 2604.20039 ↗First posted 21 Apr 2026cs.AI cs.LGcited by 0

On the map: LLM reasoning → Causal reasoning in LLMs

also in 1 more
Abstract (arXiv)

Causal discovery through experimentation and intervention is fundamental to robust problem solving. It requires not just updating beliefs within a fixed framework but revising the hypothesis space itself, a capacity current AI agents lack when evidence demands representations they have not previously constructed. We extend the blicket detector paradigm from developmental science to test this capacity in AI agents equipped with architectural scaffolding that targets hypothesis-space restructuring. Our compositional architecture has two discrete components: context graphs, which structure exploration as typed state machines, and dynamic behaviors, which monitor for evidence that the current hypothesis space is inadequate and expand it at runtime. Across 1,085 experimental trials, these components make orthogonal contributions: context graphs drive reasoning quality within the post-switch hypothesis space, accounting for 94% of the accuracy gain, while dynamic behaviors drive reasoning eligibility by detecting regime changes and preventing premature commitment to outdated hypotheses.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Causal reasoning in LLMs, one of 17 directions in LLM reasoning. Builds on 12 of its 39 references that are in the map. First posted on arXiv 21 Apr 2026.

Builds on

12 of 39 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023Causal Reasoning and Large Language Models: Opening a New Frontier for Causality28 Apr

Other references · 11

  1. 2024AFlow: Automating Agentic Workflow Generationcited by 432
  2. 2024Mechanistic Interpretability for AI Safety -- A Reviewcited by 422
  3. 2015Neural Module Networkscited by 376
  4. 2023Causal Parrots: Large Language Models May Talk Causality But Are Not Causalcited by 150
  5. 2023Can Large Language Models Infer Causation from Correlation?cited by 146
  6. 2024StateFlow: Enhancing LLM Task-Solving through State-Driven Workflowscited by 65
  7. 2024Do Large Language Models Have Compositional Ability? An Investigation into Limitations and Scalabilitycited by 40
  8. 2022Towards Understanding How Machines Can Learn Causal Overhypothesescited by 14
  9. 2024Reflection-Bench: Evaluating Epistemic Agency in Large Language Modelscited by 7
  10. 2025Language Agents Mirror Human Causal Reasoning Biases. How Can We Help Them Think Like Scientists?cited by 7
  11. 2026BabyReasoningBench: Generating Developmentally-Inspired Reasoning Tasks for Evaluating Baby Language Modelscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026ACED-Bench: Evaluating How LLM Agents Acquire and Act on Causal Evidence
  2. 2026Tiny Moves: Game-based Hypothesis Refinement
  3. 2026CausaLab: A Scalable Environment for Interactive Causal Discovery Toward AI Scientists
  4. 2026AI scientists produce results without reasoning scientifically
  5. 2026CausalSteward: An Agentic Divide-Conquer-Combine Copilot for Causal Discovery
  6. 2026CausalGame: Benchmarking Causal Thinking of LLM Agents in Games
  7. 2025Toward Causal-Visual Programming: Enhancing Agentic Reasoning in Low-Code Environments
  8. 2026The Reasoning Error About Reasoning: Why Different Types of Reasoning Require Different Representational Structures