Search papers, labs, and topics across Lattice.
1
0
3
A 4B model can rival the mathematical reasoning of models 30x its size, proving that clever training trumps brute force scaling.
