Update README.md

README.md

@@ -7,4 +7,5 @@ sdk: static
 pinned: false
 ---
 
-Edit this `README.md` markdown file to author your organization card.
+🐙 OctoThinker, led by [GAIR](https://huggingface.co/GAIR), is an initiative to explore earlier training interventions that make base models more amenable to reinforcement learning (RL) scaling.
+🎯 Our Goal: To reshape the pre-training trajectory so models scale better under RL.