Communication Dans Un Congrès Année : 2025

Hallucination-Resilient LLM-Driven Sound and Tunable Static Analysis

Résumé

We argue that soundness remains essential for LLM-driven static analysis and discuss hallucination-resilient approaches in combining LLMs with static analysis that ensure soundness while improving precision. We propose to use LLMs as a way of meta-analysis and investigate this approach in higher-order control-flow analysis, building on the abstracting abstract machine framework and delegating abstract address allocation to an LLM. Our analyzer llmaam maintains soundness regardless of LLM behavior, while adaptively tuning analysis precision. We report promising preliminary results and outline broader opportunities for sound LLMdriven analysis.

Fichier principal
Vignette du fichier
paper.pdf (571.94 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05296479 , version 1 (03-10-2025)

Licence

Identifiants

Citer

Guannan Wei, Zhuo Zhang, Caterina Urban. Hallucination-Resilient LLM-Driven Sound and Tunable Static Analysis. LMPL 2025 - 1st International Workshop on Language Models and Programming Languages, Oct 2025, Singapore, Singapore. ⟨10.1145/3759425.3763378⟩. ⟨hal-05296479⟩
288 Consultations
473 Téléchargements

Altmetric

Partager

  • More