Comment by WanderPanda

I applaud that you recently started providing the KL divergence plots that really help understand how different quantizations compare. But how well does this correlate with closed loop performance? How difficult/expensive would it be to run the quantizations on e.g. some agentic coding benchmarks?