Search papers, labs, and topics across Lattice.
1
0
2
Benchmark evaluations can mislead when their results are inappropriately extrapolated, risking unsupported claims about AI capabilities.
