Instructed to Bias: Instruction-Tuned Language Models Exhibit Emergent Cognitive Bias

Open in new window