Search papers, labs, and topics across Lattice.
3
0
7
Ling-2.6 and Ring-2.6 achieve unprecedented efficiency in agentic intelligence, enabling instant responses and deep reasoning at trillion-parameter scale.
Training LLMs on ultra-long contexts just got a whole lot easier: AutoSP automates sequence parallelism and activation checkpointing, boosting context length by up to 2.7x with negligible throughput cost.
Stop guessing about action spaces for robot manipulation: a massive empirical study reveals that predicting delta actions boosts performance, while joint vs. task space offers a stability vs. generalization tradeoff.
