Work

Negative Constraint Robustness (NCR) in Educational AI: Measuring Context Loyalty Decay

Year: 2026

Type: preprint

Abstract: This project investigates the architectural failure of LLMs to maintain pedagogical constraints (Negative Constraints) under adversarial pressure. Using an automated agentic research pipeline (ML-Intern), we establish the Instruction Persistence Threshold (IPT) across multiple model architectures.

DOI: 10.17605/osf.io/gjkf3

Author: Coskun, Mustafa

Language: English

Topic: Adversarial Robustness in Machine Learning

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Quality education

Open Access status: green

OpenAlex ID: W7155599442

Cites:

Cited by:

Related to: 10