Applied Compute(@appliedcompute):Introducing Spork, an efficient post-training recipe which increases our solve-rate on chess puzzles by 100% when compared to our best GRPO setup. It uses a GRPO warmup, splits task-specific datasets further into partitions, trains specialists, and then uses multi-teacher on-policy distillation. Like a Spork, it leverages gains by simply consolidating specialists.

Applied Compute

@appliedcompute

The best AI is built, not bought. Our platform, Applied Compute Agent Cloud, is now in private beta. Book a demo below.

参加 July 2012

24 フォロー中 8.3K ファン

Applied Compute@appliedcompute

2026.10.13 22:07

Introducing Spork, an efficient post-training recipe which increases our solve-rate on chess puzzles by 100% when compared to our best GRPO setup. It uses a GRPO warmup, splits task-specific datasets further into partitions, trains specialists, and then uses multi-teacher on-policy distillation. Like a Spork, it leverages gains by simply consolidating specialists.