Search papers, labs, and topics across Lattice.
1
0
3
Small language models can achieve surprisingly large gains in reasoning accuracy and efficiency by learning to follow explicit, structured reasoning paths guided by simple control tags during training.
