Running Featured 63 Distilling 100B+ Models 40x Faster with TRL π 63 TRL distillation for 100B+ teachers, 40x faster
allenai/tulu-3-sft-personas-instruction-following Viewer β’ Updated Nov 21, 2024 β’ 30k β’ 5.56k β’ 64
Running on CPU Upgrade Featured 3.11k The Smol Training Playbook π 3.11k The secrets to building world-class LLMs
mrlbenchmarks/global-piqa-nonparallel-v0.1 Viewer β’ Updated Oct 29, 2025 β’ 11.6k β’ 5.78k β’ 36
Running Featured 1.33k FineWeb: decanting the web for the finest text data at scale π· 1.33k Explore and download the FineWeb webβtext dataset