Michel Ma
@ma-michel.bsky.social
Reinforcement learning researcher - PhD candidate @Mila
Reposted by Michel Ma
Can we make LLMs reason effectively without a huge inference time cost?
We show a powerful approach through learning and forgetting!
Our recipe: ⬇️
We show a powerful approach through learning and forgetting!
Our recipe: ⬇️
April 23, 2025 at 10:05 PM
Can we make LLMs reason effectively without a huge inference time cost?
We show a powerful approach through learning and forgetting!
Our recipe: ⬇️
We show a powerful approach through learning and forgetting!
Our recipe: ⬇️