Compute-Optimal LLMs Provably Generalize Better With Scale

Re-assign community