Search papers, labs, and topics across Lattice.
University of Tsukuba
2
0
4
LLM agents can now remember far more, far more accurately, by "seeing" their past experiences instead of just reading about them.
Open-source MLLMs can now achieve state-of-the-art accuracy on complex tabular reasoning tasks, even outperforming models 18x their size, by explicitly penalizing visual hallucinations and shortcut guessing through process-supervised RL.
