Saved in:
Bibliographic Details
Main Authors: Ullah, Saad, Han, Mingji, Pujar, Saurabh, Pearce, Hammond, Coskun, Ayse, Stringhini, Gianluca
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2312.12575
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917731528343552
author Ullah, Saad
Han, Mingji
Pujar, Saurabh
Pearce, Hammond
Coskun, Ayse
Stringhini, Gianluca
author_facet Ullah, Saad
Han, Mingji
Pujar, Saurabh
Pearce, Hammond
Coskun, Ayse
Stringhini, Gianluca
contents Large Language Models (LLMs) have been suggested for use in automated vulnerability repair, but benchmarks showing they can consistently identify security-related bugs are lacking. We thus develop SecLLMHolmes, a fully automated evaluation framework that performs the most detailed investigation to date on whether LLMs can reliably identify and reason about security-related bugs. We construct a set of 228 code scenarios and analyze eight of the most capable LLMs across eight different investigative dimensions using our framework. Our evaluation shows LLMs provide non-deterministic responses, incorrect and unfaithful reasoning, and perform poorly in real-world scenarios. Most importantly, our findings reveal significant non-robustness in even the most advanced models like `PaLM2' and `GPT-4': by merely changing function or variable names, or by the addition of library functions in the source code, these models can yield incorrect answers in 26% and 17% of cases, respectively. These findings demonstrate that further LLM advances are needed before LLMs can be used as general purpose security assistants.
format Preprint
id arxiv_https___arxiv_org_abs_2312_12575
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle LLMs Cannot Reliably Identify and Reason About Security Vulnerabilities (Yet?): A Comprehensive Evaluation, Framework, and Benchmarks
Ullah, Saad
Han, Mingji
Pujar, Saurabh
Pearce, Hammond
Coskun, Ayse
Stringhini, Gianluca
Cryptography and Security
Large Language Models (LLMs) have been suggested for use in automated vulnerability repair, but benchmarks showing they can consistently identify security-related bugs are lacking. We thus develop SecLLMHolmes, a fully automated evaluation framework that performs the most detailed investigation to date on whether LLMs can reliably identify and reason about security-related bugs. We construct a set of 228 code scenarios and analyze eight of the most capable LLMs across eight different investigative dimensions using our framework. Our evaluation shows LLMs provide non-deterministic responses, incorrect and unfaithful reasoning, and perform poorly in real-world scenarios. Most importantly, our findings reveal significant non-robustness in even the most advanced models like `PaLM2' and `GPT-4': by merely changing function or variable names, or by the addition of library functions in the source code, these models can yield incorrect answers in 26% and 17% of cases, respectively. These findings demonstrate that further LLM advances are needed before LLMs can be used as general purpose security assistants.
title LLMs Cannot Reliably Identify and Reason About Security Vulnerabilities (Yet?): A Comprehensive Evaluation, Framework, and Benchmarks
topic Cryptography and Security
url https://arxiv.org/abs/2312.12575