Zhao Xin

Zhao Xin is a dedicated researcher at the Language Technology Group (LLMC) within National Institute of Informatics, focusing on Natural Language Processing and Knowledge Representation. His academic journey includes a Doctoral Courser at The University of Tokyo, a Master Courser at Nara Institute of Science and Technology, and a Bachelor’s in Japanese at Xi'an Jiaotong University.

• Education:
  • Doctoral Courser (2023–2026), Information Science and Technology, The University of Tokyo
  • Master Courser (2018–2020), Computer Science, Nara Institute of Science and Technology
  • Bachelor Courser (2013–2017), Foreign Languages and Literatures, Xi'an Jiaotong University

His research interests span Natural Language Processing, Domain Adaptation, Cross-lingual Transfer, Model Interpretability, and Semantic Search. Recent projects include analyzing neuron-level controllability in language models and cross-lingual knowledge transfer for Japanese NLP tasks.

Zhao’s publications (2024–2025) emphasize fact knowledge evaluation, neuron activation analysis, and cross-lingual entity alignment. He received the Young Investigator Award for his work on multilingual knowledge tracing.

Proficient in Japanese and English, Zhao contributes to tools like llm-jp-eval and Megatron-LM, with expertise in LLM pre-training, domain adaptation, and Python/Machine Learning frameworks.