Anchors in the Machine: Behavioral and Attributional Evidence of Anchoring Bias in LLMs

Open in new window