Mercor is assembling an expert panel to red-team frontier AI models, focusing on misuse potential and safe responses. You will craft challenging prompts in benign, dual-use, and adversarial categories.
You will also evaluate model outputs against a policy standard and write the reference answer with detailed reasoning, drawing on real hardware experience and rigorous technical writing.
#J-18808-Ljbffr…
