The Threat of Package Hallucinations

In this episode of ConversingLabs, host Paul Roberts interviews Major Joe Spracklen, a PhD student at the University of Texas at San Antonio, who recently published a paper with his peers regarding the threat posed to software supply chains caused by code-generating Large Language Models (LLMs). The paper, “We Have a Package for You! A Comprehensive Analysis of Package Hallucinations by Code Generating LLMs” (PDF), discusses how the rise of these LLMs can create package hallucinations that arise from fact-conflicting errors – representing a novel form of package confusion attack.

There are no comments yet.

Please log in to write the first comment.