REER-PT: Reverse-Engineered Reasoning for Perplexity-Guided Pre-training Data Augmentation
Haoran Que, Jiajun Shi, Ting Huang, Renming Pang, et al.
arXiv preprint · First author
AI RESEARCHER
Researcher · Peking University
I study large language models, with a focus on model evaluation. We aim to build benchmarks that reflect real-world data distributions and evaluate models from pre-training through post-training.
Haoran Que, Jiajun Shi, Ting Huang, Renming Pang, et al.
arXiv preprint · First author
Haoran Que, Wenge Rong
ACL · First author
Haoran Que, Jiaheng Liu, Ge Zhang, Chenchen Zhang, et al.
NeurIPS · Co-first author
Haoran Que, Feiyu Duan, Liqun He, Yutao Mou, et al.
arXiv preprint · Co-first author
Zekun Moore Wang, Zhongyuan Peng, Haoran Que, Jiaheng Liu, et al.
Findings of ACL · Co-first author
Jiaheng Liu, Zehao Ni, Haoran Que, Tao Sun, et al.
NeurIPS · Datasets and Benchmarks · Co-first author