Search papers, labs, and topics across Lattice.
1
0
3
5
Current research agent benchmarks miss crucial aspects of real-world research, like multimodal reasoning and iterative refinement, which MiroEval now captures.
