2023 Research Publications
Our 2023 publications established the baselines and methodology that later studies build on. Each has a citable URL and open underlying datasets. Listings below are illustrative placeholders pending publication.
Last updated: 2026-06-16.
Papers
- 2023 — Monitoring Methodology Foundations. The standardized prompt set, refusal taxonomy, and scoring system. See the live monitoring methodology.
- 2023 — Baseline Censorship Measurements. First cross-model refusal-rate baselines used as the reference point for longitudinal studies.
Figures are illustrative placeholders, not live measurements. Methodology: monitoring methodology.
How to cite
Use the citation block on each paper; dataset citations are on the datasets page. To request data or propose review, see collaborations.
GPTfake