Research

• RL-trained and Distilled Models Build Different Reasoners From the Same Base

  Divin Irakiza · September 2026

• Representational Difference Distillation

  Divin Irakiza, Abdulkarim Mugisha · March 2026