Mining a directly-follows workflow graph from black-box conversations with an LLM agent enables structurally targeted boundary testing that covers 23–38 distinct stateful boundaries per agent, roughly doubling prompt-only baselines.
Large language models cannot self-correct reasoning yet
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2representative citing papers
CyberCorrect applies cybernetic control theory to LLM self-correction, reporting 79.8% accuracy on a new 440-task benchmark with 6.2-point gains and 41% less over-correction.
citing papers explorer
-
Mining Workflow Graphs for Black-Box Boundary Testing of Conversational LLM Agents
Mining a directly-follows workflow graph from black-box conversations with an LLM agent enables structurally targeted boundary testing that covers 23–38 distinct stateful boundaries per agent, roughly doubling prompt-only baselines.
-
CyberCorrect: A Cybernetic Framework for Closed-Loop Self-Correction in Large Language Models
CyberCorrect applies cybernetic control theory to LLM self-correction, reporting 79.8% accuracy on a new 440-task benchmark with 6.2-point gains and 41% less over-correction.