BenchmarkHybridCodeAuthorship基准
HybridCodeAuthorship
An external benchmark dataset for evaluating the detection of authorship in human-machine collaborative code. It has no training split and focuses on the task of tracing the provenance of code that mixes AI and human contributions.
Timeline (last 90 days)
Oct 21
On the HybridCodeAuthorship external benchmark (no training set), the method achieved a zero-shot line-level F1 of 62.5, exceeding the best line-level result of 56 reported in the original paper.
Unverified50%