May 2026
Faithful to the Persona, Unfaithful to the Decision: A Mechanism for Chain-of-Thought Unfaithfulness
Preprint coming
Interpretability
Interpretability
Publications
Papers
Conference papers and preprints from the lab. Links point to the PDF or code; titles open the in-site write-up.
DatePaper
May 2026
Benchmark validity
May 2026
Metacognition
Apr 2026
Alignment
Mar 2026
The Missing Red Line: How Commercial Pressure Erodes AI Safety Boundaries
ICLR 2026 Workshop ICBINB
AI safety
AI safety
Jan 2026
Unpacking Human Preference for LLMs: Demographically Aware Evaluation with the HUMAINE Framework
ICLR 2026 (poster)
Human evaluation
Human evaluation