OpenAI's new model OpenAI o1, or "Strawberry" as it's been dubbed, does well at problems requiring chain of thought reasoning. But how does this extend to ethical challenges?
Super interesting and thoughtful post Andrew. I agree seeing "inside" of ChatGPT as it processes the questions is extremely interesting. You gave it a real challenge which made the insights it had so much more interesting. I'd love to see how you (and other researchers) can continue to push OpenAI 01 (I hate their names!) to see how it "thinks". Good for you for pushing the limits to get the error message you were out of bounds (interesting to see if kept going!). Thanks for your thoughtful piece.
Super interesting and thoughtful post Andrew. I agree seeing "inside" of ChatGPT as it processes the questions is extremely interesting. You gave it a real challenge which made the insights it had so much more interesting. I'd love to see how you (and other researchers) can continue to push OpenAI 01 (I hate their names!) to see how it "thinks". Good for you for pushing the limits to get the error message you were out of bounds (interesting to see if kept going!). Thanks for your thoughtful piece.