Search papers, labs, and topics across Lattice.
2
0
3
2
Existing models miss critical visual evidence in metaphor understanding, but M$^3$R-Reasoner closes the gap, outperforming larger models in both accuracy and justification metrics.
LLMs fall short of clinician performance in psychiatric evaluations, trailing by over 37 percentage points in objective competence.
