AI safety researcher

Rohin Shah

Google Scholar dblp

Last 12 months4 papers

Defenses 1

AI safety papers from the last year

On this site: 15 papers

August 2026

Debate Training Reduces Reward Hacking in RLAIF
Aug 18DefensesarXiv ↗

June 2026

How Transparent is DiffusionGemma?
Jun 18arXiv ↗

May 2026

Realistic honeypot evaluations for scheming propensity
May 28arXiv ↗

January 2026

Building Production-Ready Probes For Gemini
Jan 16arXiv ↗

Authorship follows dblp and the author's own Google Scholar profile; only AI safety papers on this site are listed.