Days after former employees said the company was being too reckless with its technology, OpenAI released a research paper on a method for reverse engineering the workings of AI models.
https://www.wired.com/story/openai-offers-a-peek-inside-the-guts-of-chatgpt/