Maheep Chaudhary

Co­er­cion and De­cep­tion in AI-to-AI Management

Would your AI travel agent book a bul­lfight? Test­ing whether agents con­sider an­i­mal welfare with­out be­ing prompted

Aware­ness Jailbreak­ing: Re­veal­ing True Align­ment in Eval­u­a­tion-Aware Models

• Maheep Chaudhary 19 Dec 2025 7:52 UTC

  3 points

  0

  on: Ma­heep Chaud­hary’s Shortform

  I just published my first paper as an article on less wrong, on the topic of evaluation awareness.
  
  TL;DR: We found “evaluation awareness” increases as the model size increases. Evaluation awareness is the ability of the model to differentiate if it is being evaluated or is deployed.
  
  This will have serious concern if not countered. Your feedback would be highly appreciated on the article.

Eval­u­a­tion Aware­ness Scales Pre­dictably in Open-Weights Large Lan­guage Models