8 papers
Falsehood and Impossibility Are Different Directions in an AI's Representation of Language
Yoon Pyo Lee
Language can describe states of affairs that are false and states of affairs that could not be the case at all. Whether an AI model internally distinguishes these failures remains…
Multimodal Language Models Benchmarked Against the NRC Reactor Operator Licensing Examination: Fine-Tuning and Retrieval Strategies
Isak Hwang, Yoon Pyo Lee, Syed Bahauddin Alam
Competence claims for a language model in a safety-critical domain are credible when measured against a standard the domain already enforces. We evaluate an open-weight 31-billion-…
Semantic Intent Fragmentation: A Single-Shot Compositional Attack on Multi-Agent AI Pipelines
Tanzim Ahad, Ismail Hossain, Md Jahangir Alam +4
We introduce Semantic Intent Fragmentation (SIF), an attack class against LLM orchestration systems where a single, legitimately phrased request causes an orchestrator to decompose…
When Safety Geometry Collapses: Fine-Tuning Vulnerabilities in Agentic Guard Models
Ismail Hossain, Sai Puppala, Jannatul Ferdaus +4
A guard model fine-tuned on entirely benign data can lose all safety alignment -- not through adversarial manipulation, but through standard domain specialization. We demonstrate t…
Agent-Fence: Mapping Security Vulnerabilities Across Deep Research Agents
Sai Puppala, Ismail Hossain, Md Jahangir Alam +5
Large language models are increasingly deployed as *deep agents* that plan, maintain persistent state, and invoke external tools, shifting safety failures from unsafe text to unsaf…
Scientific feedback reveals productive regimes beyond training conventions
Yoon Pyo Lee, Yoonpyo Lee, Samrendra Roy +2
Human-imposed design conventions can restrict the alternatives represented in scientific training data and shape the search preferences learned by generative models. We show that s…
Agentic Physical AI toward a Domain-Specific Foundation Model for Energy Systems: A Case Study on Nuclear Reactor Control
Yoon Pyo Lee, Samrendra Roy, Kazuma Kobayashi +5
The prevailing paradigm in AI for physical systems: scaling general-purpose foundation models toward universal multimodal reasoning, confronts a barrier at the control interface. F…
Mechanistic Interpretability of LoRA-Adapted Language Models for Nuclear Reactor Safety Applications
Yoon Pyo Lee
The integration of Large Language Models (LLMs) into safety-critical domains, such as nuclear engineering, necessitates a deep understanding of their internal reasoning processes.…