RPDR: A Round-trip Prediction-Based Data Augmentation Framework for Long-Tail Question Answering
arXiv:2602.17366v1 Announce Type: new Abstract: Long-tail question answering presents significant challenges for large language models (LLMs) due to their limited ability to acquire and accurately …
Yiming Zhang, Siyue Zhang, Junbo Zhao, Chen Zhao
9 views
