You can extract the thinking process from closed LLMs
Researchers figured out how to steal the internal reasoning steps from proprietary models like GPT-4o and Claude through API timing attacks.
Read the note 1 min read
From the notebook
side-channel-attacks.
1 note on this topic.