Manglam Kartik

Capability Is Not Propensity: Measuring Pressure-Robust Cooperative Behavior in Civic LLM Agents

This work argues that Cooperative AI evaluations should separate what models can do under benign instructions from what they tend to do under realistic civic pressure, and introduces DiffCoop-Civic, a 10-scenario pilot evaluation suite spanning preference understanding, evidence and persuasion, commitment design, asymm...