Search papers, labs, and topics across Lattice.
University of Tennessee
1
0
2
LLM-generated unit tests can detect faults nearly four times more effectively than human-written tests, challenging the assumption that coverage is a reliable measure of test quality.