Last released Oct 22, 2026
Distill an expensive LLM classifier into a tiny model: label with a teacher LLM, train compact students, benchmark cost/accuracy/latency, serve on CPU.