Implicit Patterns in LLM-Based Binary Analysis

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Qiang, Zhang, XiangRui, Wang, Haining
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917353593241600
author Li, Qiang
Zhang, XiangRui
Wang, Haining
author_facet Li, Qiang
Zhang, XiangRui
Wang, Haining
contents Binary vulnerability analysis is increasingly performed by LLM-based agents in an iterative, multi-pass manner, with the model as the core decision-maker. However, how such systems organize exploration over hundreds of reasoning steps remains poorly understood, due to limited context windows and implicit token-level behaviors. We present the first large-scale, trace-level study showing that multi-pass LLM reasoning gives rise to structured, token-level implicit patterns. Analyzing 521 binaries with 99,563 reasoning steps, we identify four dominant patterns: early pruning, path-dependent lock-in, targeted backtracking, and knowledge-guided prioritization that emerge implicitly from reasoning traces. These token-level implicit patterns serve as an abstraction of LLM reasoning: instead of explicit control-flow or predefined heuristics, exploration is organized through implicit decisions regulating path selection, commitment, and revision. Our analysis shows these patterns form a stable, structured system with distinct temporal roles and measurable characteristics. Our results provide the first systematic characterization of LLM-driven binary analysis and a foundation for more reliable analysis systems.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19138
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Implicit Patterns in LLM-Based Binary Analysis
Li, Qiang
Zhang, XiangRui
Wang, Haining
Artificial Intelligence
Cryptography and Security
Software Engineering
Binary vulnerability analysis is increasingly performed by LLM-based agents in an iterative, multi-pass manner, with the model as the core decision-maker. However, how such systems organize exploration over hundreds of reasoning steps remains poorly understood, due to limited context windows and implicit token-level behaviors. We present the first large-scale, trace-level study showing that multi-pass LLM reasoning gives rise to structured, token-level implicit patterns. Analyzing 521 binaries with 99,563 reasoning steps, we identify four dominant patterns: early pruning, path-dependent lock-in, targeted backtracking, and knowledge-guided prioritization that emerge implicitly from reasoning traces. These token-level implicit patterns serve as an abstraction of LLM reasoning: instead of explicit control-flow or predefined heuristics, exploration is organized through implicit decisions regulating path selection, commitment, and revision. Our analysis shows these patterns form a stable, structured system with distinct temporal roles and measurable characteristics. Our results provide the first systematic characterization of LLM-driven binary analysis and a foundation for more reliable analysis systems.
title Implicit Patterns in LLM-Based Binary Analysis
topic Artificial Intelligence
Cryptography and Security
Software Engineering
url https://arxiv.org/abs/2603.19138