Search papers, labs, and topics across Lattice.
Affiliation:
2
0
5
Extra test-time "thinking" cannot recover missing parametric facts, but teaching small reasoning models to recognize their own knowledge boundaries and selectively query larger models lets a 4B model outperform a 14B baseline at nearly a third of the serving cost.
Robot swarms can autonomously self-assemble power networks that are 65% more efficient than traditional methods, even in chaotic environments.
