Search papers, labs, and topics across Lattice.
3
38
8
9
Larger models learn more not just because of increased capacity, but because they experience less interference during training, allowing them to retain rare and complex tasks that smaller models forget.
Even massive LLMs like GPT-5.4 hit a surprisingly low "depth ceiling" of just 5-7 steps when discovering multi-step planning strategies latently, suggesting that complex reasoning may require explicit training or externalization.
Command A shows how to build an enterprise-grade LLM that balances performance, efficiency, and multilingual capabilities using decentralized training and model merging.
