Unveiling Downstream Performance Scaling of LLMs: A Clustering-Based Perspective

Open in new window