Model1h ago
Reasoning Models Vulnerable To Denial-of-Service Attacks
Internal monologues in reasoning LLMs create a new security flaw. Attackers can trigger excessively long reasoning chains, forcing models to overthink and slowing system performance to a crawl. This vulnerability targets the step-by-step logic used in coding and math tasks. Developers must now balance complex reasoning with strict output constraints.