Search papers, labs, and topics across Lattice.
3
0
4
Graph-based code representations, largely unexplored in automated patch correctness assessment, crush sequence- and heuristic-based methods, achieving 82.6% accuracy in predicting patch correctness.
Forget finetuning: carefully chosen context can boost LLM performance on software engineering tasks by up to 33%, and CL4SE provides the benchmark to prove it.
LLM-powered test generation can finally achieve meaningful coverage in deep learning libraries, thanks to a novel agent-driven framework that iteratively refines tests based on constraint validation.
