From the lab
Papers, replications and technical notes from the lab.
Our first release: a small reasoning model built on LLaMA-3.1 8B that breaks down its thinking into transparent, structured steps.