🗣️ Foundation Models

Data-Centric Learning / Data Selection

Pre-training Data Selection

• MuRating: A High Quality Data Selecting Approach to Multilingual Large Language Model Pretraining
  Zhixun Chen, Ping Guo, Wenhan Han, Yifan Zhang, Binbin Liu, Haobin Lin, Fengze Liu, Yan Zhao, Bingni Zhang, Taifeng Wang, Yin Zheng, Trevor Cohn, Meng Fang | NeurIPS 2025
  Multilingual pre-training data quality scoring and selection: selects multilingual subsets from web data for pretraining LLaMA-style models.
• ParaRater: Enhancing Cross-Lingual Transfer in LLMs with Meta-Learning
  Wenhan Han, Binbin Liu, Zhixun Chen, Yifan Zhang, Mykola Pechenizkiy, Meng Fang, Yin Zheng | EMNLP 2026
  Parallel-corpus selection for cross-lingual transfer, using meta-learned raters to identify the most valuable parallel data.

Post-training Data Selection

• DOSE: Data Selection for Multi-Modal LLMs via Off-the-Shelf Models
  Biao Wu, Yiwu Zhong, Meng Fang, Ling Chen | ACL Findings 2026
  Multimodal instruction-tuning data selection: scores text quality and image–text alignment, then applies diversity-aware sampling; evaluated on LLaVA vision-language instruction tuning.