Search papers, labs, and topics across Lattice.
1
0
2
LLM-generated rubrics can nearly match human evaluation standards, but they often miss the mark with excessive detail and scoring bias.
