Research
Independent research projects on contextual safety, human-AI interaction, and signal integrity.
All papers are publicly accessible via Zenodo or SSRN. Direct links to original versions only.
CXC-7
The Seven Core Dimensions of Conversational Context
A seven-dimensional taxonomy for analyzing conversational context risk as a multi-dimensional structure. Grounded in attachment theory, framing effects, and conversational AI safety research.
v1.1.0 · 2025
CXOD-7
CXOD-7 and Coh(G): A Contextual Offense and Defense Evaluation Framework for Large Language Models
System-side contextual operation dimensions with Contextual Coherence metric Coh(G). Features 7×7 offense-defense matrix experiments across five major LLMs.
USCH
User-Side Contextual Hallucination in Human-AI Interaction
A non-clinical construct describing how users progressively form structurally deviated understandings of AI systems through prolonged interaction. Proposes six-stage formation process and fourteen observable phenomena.
Preprint · January 2026
USCI
User-Side Contextual Interaction Assessment Method
Pre-empirical methodology with four-axis scoring (FR, CA, SR, SA) for post-interaction user-side contextual risk evaluation.
v1.0.0 · 2026
—
Upcoming Research
Details will be announced upon publication.
—
—
Upcoming Research
Details will be announced upon publication.
—
Complete research architecture, framework details, and academic communication.