Search papers, labs, and topics across Lattice.
3
0
6
12
CRPO effectively mitigates exposure bias in self-distillation, leading to superior performance in complex reasoning tasks.
WrapFlow revolutionizes irregular time series forecasting by eliminating the need for complex numerical solvers while achieving state-of-the-art accuracy.
Process evaluations reveal hidden failures in LLM reasoning, showing that lucky successes can mask critical deficiencies in agent performance.