Context graphs and dynamic behaviors provide orthogonal contributions to causal reasoning in LLM agents, with graphs improving accuracy within a hypothesis space and dynamic behaviors enabling detection of regime changes to restructure the space.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Separable Pathways for Causal Reasoning: How Architectural Scaffolding Enables Hypothesis-Space Restructuring in LLM Agents
Context graphs and dynamic behaviors provide orthogonal contributions to causal reasoning in LLM agents, with graphs improving accuracy within a hypothesis space and dynamic behaviors enabling detection of regime changes to restructure the space.