How to separate reasoning traces into distinct reasoning steps

[zhuyuzy]

Hello team, thank you for your nice work.

I'm wandering how to reproduce your reasoning trace separating like your paper said

We use Llama-3.3-70B-Instruct (Dubey et al., 2024) to separate each reasoning trace into distinct reasoning steps, with boundaries determined by occurrences of backtracking, selfvalidation, reflection, or other breaks from a linear sequence of thoughts.

Could you share your prompt?