Search papers, labs, and topics across Lattice.
2
0
4
MLLMs can "think" with images, but their actions often don't match their reasoning, and this paper solves that with a new training method that forces them to explain what they see.
Verification is the secret sauce: an 8B parameter research agent, fortified with verification mechanisms, can now rival or surpass the performance of 30B parameter agents while drastically reducing computational cost.
