Search papers, labs, and topics across Lattice.
3
3
5
3
Current VLMs show alarming performance drops in long-context scenarios, revealing that they may be overfitting to benchmark artifacts rather than demonstrating true understanding.
Evolved agents can learn to dynamically coordinate multiple retrieval strategies, leading to a remarkable 19.6-point performance boost in multimodal document reasoning.
Frontier-level multimodal reasoning is now within reach for organizations with limited infrastructure, thanks to a 15B parameter model that rivals much larger models through clever training design, not brute force scaling.
