Large Language Models Still Exhibit Bias in Long Text

Wonje Jeung, Dongjae Jeon, Ashkan Yousefpour +1

Existing fairness benchmarks for large language models (LLMs) primarily focus on simple tasks, such as multiple-choice questions, overlooking biases that may arise in more complex…