What Benchmarks Don't Measure: The Case for Evaluating Abstention Competence in Autonomous Agents

본문 미리보기

arXiv:2606.02965v1 Announce Type: new Abstract: Benchmarks for autonomous agents measure whether agents complete tasks, yet this framing is systematically blind to whether an agent should have proceeded at all. Agents trained under human-feedback objectives develop a structural tendency to proceed even when they lack the inputs, evidence, or authorization to act safely, a disposition we term compliance bias, because both the reward signal and the benchmark scoring regime treat proceeding as the