Search papers, labs, and topics across Lattice.
1
0
2
Coreference benchmarks may be overstating language models' NLU abilities, as even small changes to evaluation contexts reveal a failure to generalize.
