Why it matters
Chinchilla changed how models are trained. Understanding shapes model design.
Advertisement
The architecture
Fit L(N, D) with power law.
Optimize under compute constraint.
Optimal N proportional to D.
Advertisement
How it works end to end
Chinchilla optimum: N proportional to D^0.5 approximately.
Approx: 20 tokens per param.
Modern practice: over-train for inference efficiency (Llama 3 = 1000+ tokens per param).