[BSL2025] Backdooring LLM and bypassing HuggingFace malware scanners
Davide Cioccia explores [BSL2025] Backdooring LLM and bypassing HuggingFace malware scanners.
Summary
Backdooring large language models and bypassing HuggingFace malware scanners.
Related talks
Briefings·37 min
Tien-Chih Lin · Wei Chieh Chao · Zhao-Min Chen
IDA Pro feat. MCP (Model Context Protocol) is truly amazing.
reverse-engineeringmalware-analysisctf+3
Black Hat USA·36 min
Sahar Abdelnabi · Christoph Endres · Mario Fritz · Kai Greshake · Shailesh Mishra
Engineers explore architecture and security mechanisms for Compromising LLMs: The Advent of AI Malware.
malware-analysiscompromisingllm+1
Briefings·41 min
Estelle Ruellan · Olivier Bilodeau
Threat analysts uncover operational security blunders and digital breadcrumbs left by state-sponsored threat actors during cyber heists.
social-engineeringthreat-intelligenceai+3