Fetching the paper…

AI-Slop to AI-Polish? Aligning Language Models through Edit-Based Writing Rewards and Test-time Computation · Around