Secure By Dezign
Context Window Poisoning: Long-Co...
AI
Context Window Poisoning: Long-Context LLM Attacks in 128K+ Token Models
AI

Secure By Dezign by Pax

Mar 2, 2026
04:37
Episode notes

How attackers weaponize 128K+ token context windows to hide malicious instructions in long documents, evade

Keywords
llmsecurityapiaihackers