EchoCoT: Extracting Hidden Chain-of-Thought from Large Reasoning Models
Quick summary
arXiv:2608.20055v1 Announce Type: cross Abstract: Hidden chain-of-thought (CoT) traces, especially those from frontier proprietary large reasoning models (LRMs), are valuable model assets. Yet whether these hidden CoTs can be directly extracted from black-box models remains largely unexplored. In this work, we systematically study whether hidden CoTs can be extracted near-verbatim from black-box LRMs through API interactions. We identify a previously overlooked reasoning replay surface between tool calls and develop EchoCoT, a multi-step attack that iteratively extracts hidden CoTs using API-r
Key takeaways
- arXiv:2608.20055v1 Announce Type: cross Abstract: Hidden chain-of-thought (CoT) traces, especially those from frontier proprietary large reasoning models (LRMs), are valuable model assets.
- Yet whether these hidden CoTs can be directly extracted from black-box models remains largely unexplored.
- In this work, we systematically study whether hidden CoTs can be extracted near-verbatim from black-box LRMs through API interactions.
Why it matters
“EchoCoT: Extracting Hidden Chain-of-Thought from Large Reasoning Models” should be evaluated beyond branding and benchmark scores. Its practical importance will emerge in task accuracy, latency, unit cost, safety and integration with real workflows.
Member comments