@relizarov ("Roman Elizarov") replied to a tweet by @relizarov:
If you've followed the news on the capabilities of modern LLMs, you might have noticed that they usually give better answers to complex questions when explicitly asked to write down a step-by-step solution. /7