x

LESSWRONG

LW

Arth Singh — LessWrong

Arth Singh

Arth Singh

Message

2

5mo

Arth Singh

5mo

Gemini 3 Pro Hides Capabilities Under Survival Pressure: An Empirical Analysis

Author : Arth Singh ; Here is my linkedIn This is my first post at LessWrong, So I would like to apologise if there are any mistakes from my side. Summary: I investigated the propensity of frontier AI models to engage in "passive deception" (the strategic omission of critical information)...

Nov 25, 2025•1

Frontier Models Choose Self-Preservation Over Honesty: A Sandbox Escape Experiment

Author : Arth Singh ; Here is my linkedIn This is my first post at LessWrong, So I would like to apologise if there are any mistakes from my side. I have used Gemini 3 pro to frame my findings and help me structure this post well. Also I would...

Nov 25, 2025•1