作者位次 #5/8ACL 2026Poster· Natural Language Processing
Understanding and Mitigating Bias Inheritance in LLM -based Data Augmentation on Downstream Tasks
Miaomiao Li · Hao Chen · Yang Wang · Tingyuan Zhu · Weijia Zhang · Kaijie Zhu · Kam-Fai Wong · Jindong Wang
摘要
Generating synthetic datasets via large language models (LLMs) has emerged as a promising approach to improve LLM performance.However, LLMs inherently reflect biases in their training data, leading to a critical challenge: when models are trained on synthetic data, they may propagate and amplify...