From LLMs to Randomness: Analyzing Program Input Efficacy With Resource and Language Metrics

From LLMs to Randomness: Analyzing Program Input Efficacy With Resource and Language Metrics

Security-focused program testing typically focuses on crash detection and code coverage while overlooking additional system behaviors that can impact program confidentiality and availability. To address this gap, we propose a statistical framework that combines embedding-based anomaly detection, res...

Full description

Saved in:
Bibliographic Details
Main Authors: Gavin Black, Eric Yocam, Varghese Mathew Vaidyan, Gurcan Comert, Yong Wang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Software profiling
program behavior analysis
fuzzing techniques
resource usage metrics
large language models
anomaly detection
Online Access:https://ieeexplore.ieee.org/document/11006641/
Tags: Add Tag
No Tags, Be the first to tag this record!

Internet

https://ieeexplore.ieee.org/document/11006641/

Similar Items