一作ACL 2026Poster· Natural Language Processing
Observations and Remedies for Large Language Model Bias in Self-Consuming Performative Loop
Yaxuan Wang · Zhongteng Cai · Yujia Bao · Xueru Zhang · Yang Liu
摘要
The rapid advancement of large language models (LLMs) has led to growing interest in using synthetic data to train future models. However, this creates a self-consuming retraining loop, where models are trained on their own outputs and may cause performance drops and induce emerging biases. In re...