2 papers
cs.CR2026
Inherited Circuits, Learned Semantics: How Fine-Tuning Creates Evasion Vulnerabilities Invisible to Standard Evaluation
Ryan Fetterman
LLMs fine-tuned for security classification are usually evaluated on held-out examples from the same distribution as their training data. We show that this can miss vulnerabilities…
cs.CY2025
Toward Quantitative Modeling of Cybersecurity Risks Due to AI Misuse
Steve Barrett, Malcolm Murray, Otter Quarks +17
Advanced AI systems offer substantial benefits but also introduce risks. In 2025, AI-enabled cyber offense has emerged as a concrete example. This technical report applies a quanti…