Search papers, labs, and topics across Lattice.
3
0
6
14
A lightweight 6B model, when harnessed within the GEMS agent framework, leapfrogs state-of-the-art models in multimodal generation, suggesting architectural innovations in agents can compensate for raw parameter count.
Multimodal agents can now continually improve their tool use and orchestration in open-ended settings without parameter updates, thanks to a novel dual-stream framework that learns from both past experiences and structured skills.
LLMs can be taught to "think longer" and explore more diverse reasoning paths in-context via a simple length-incentivized reward, leading to improved generalization.
