About this document

Perplexity by smriti.rawat145 is a document available to read on EtoBox.

This document investigates the efficacy of the perplexity-based method for detecting LLM-generated code (LLMgCode) compared to feature- and pre-training-based methods. The study reveals that while perplexity excels in generalization capability, it has limitations in detection accuracy and speed, particularly for high-level programming languages. The findings provide insights for improving perplexity and its practical applications in software development.

Author

smriti.rawat145

Language

EN