PulseAugur
LIVE 10:12:50
research · [2 sources] ·
0
research

FunFuzz framework uses LLMs to improve compiler fuzzing efficiency

By PulseAugur Editorial · Summary by gemini-2.5-flash-lite from 2 sources

Researchers have developed FunFuzz, a novel evolutionary fuzzing framework that leverages Large Language Models (LLMs) to generate structured inputs for software testing. This framework addresses the sensitivity of LLM-driven fuzzing to prompt initialization by employing a multi-island approach with parallel searches and periodic migration of promising candidates. FunFuzz adapts prompts using feedback-guided selection and prioritizes candidates based on compiler coverage and failure signals, demonstrating improved coverage and discovery of unique failure-triggering inputs in tests on GCC and Clang. AI

Summary written by gemini-2.5-flash-lite from 2 sources. How we write summaries →

IMPACT Enhances software testing efficiency by improving LLM-driven fuzzing techniques for compiler development.

RANK_REASON The cluster contains an academic paper detailing a new framework for LLM-powered fuzzing.

Read on arXiv cs.CL →

COVERAGE [2]

  1. arXiv cs.CL TIER_1 Deutsch(DE) · Mario Rodr\'iguez B\'ejar, B. Romera-Paredes, Jose L. Hern\'andez-Ramos ·

    FunFuzz: An LLM-Powered Evolutionary Fuzzing Framework

    arXiv:2605.02789v1 Announce Type: cross Abstract: Modern fuzzers increasingly use Large Language Models (LLMs) to generate structured inputs, but LLM-driven fuzzing is sensitive to prompt initialization and sampling variance, which can reduce exploration efficiency and lead to re…

  2. arXiv cs.CL TIER_1 Deutsch(DE) · Jose L. Hernández-Ramos ·

    FunFuzz: An LLM-Powered Evolutionary Fuzzing Framework

    Modern fuzzers increasingly use Large Language Models (LLMs) to generate structured inputs, but LLM-driven fuzzing is sensitive to prompt initialization and sampling variance, which can reduce exploration efficiency and lead to redundant inputs. We present FunFuzz, a multi-island…

RELATED ENTITIES

RELATED TOPICS