Search papers, labs, and topics across Lattice.
1
0
3
LLMs can often achieve the same accuracy with significantly shorter self-explanations, suggesting that current chain-of-thought reasoning is unnecessarily verbose.
