Search papers, labs, and topics across Lattice.
3
0
5
2
Intern-S2-Preview-397B not only excels in multimodal scientific reasoning but also enhances biological instruction performance without altering its foundational architecture.
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.
