This experiment helps show the power of prompting reasoning in visualized steps, not a comparison to or full replication of o1, which uses different techniques. OpenAI's o1 is instead trained with ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results