Running Featured 45 Distilling 100B+ Models 40x Faster with TRL 📝 45 TRL distillation for 100B+ teachers, 40x faster
Running Featured 70 QED-Nano: Teaching a Tiny Model to Prove Hard Theorems 📝 70 Who needs 1T parameters? Olympiad proofs with a 4B model
Running on CPU Upgrade Featured 3.1k The Smol Training Playbook 📚 3.1k The secrets to building world-class LLMs
Running 61 Bringing paper to life: A modern template for scientific writing 📝 61 Explore a scientific article with interactive visualizations