Tracing Eval-Awareness Emergence Through Training of OLMo 3

TL;DR Recent work from Goodfire & UK AISI – Verbalized Eval Awareness Inflates Measured Safety – shows that newer open-weight models verbalize evaluation-awareness (VEA) more often, and that this inflates measured safety. Between OLMo-3-32B-Think and OLMo-3.1-32B-Think – identical base, SFT, DPO, an

Related Stories

📰

McDonald's Deploys AI-Powered Drive-Thru System with Google's Expertise

AI Fuels Rise of Micro-SaaS Mavericks in Indian Startup Ecosystem

Uncovr raises $7M for AI that writes the operative report before the surgeon leaves the room

Waymo Simulates Human Driver Behavior in Unpredictable Road Scenarios