Research

研究

  • Home
  • 研究業績
  • LLM / NLP

    • 研究業績

    カテゴリー

    研究領域

    年

    • 国際会議
      LLM / NLP

      GraphCheck: Breaking Long-Term Text Barriers with Extracted Knowledge Graph-Powered Fact-Checking

      [paper] [slide] [poster]

      Yingjian Chen, Haoran Liu, Yinhong Liu, Rui Yang, Han Yuan, Yanran Fu, Pengyuan Zhou, Qingyu Chen, James Caverlee, Irene Li

      The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

    • 国際会議
      LLM / NLP

      Inconsistent Tokenizations Cause Language Models to be Perplexed by Japanese Grammar

      [paper] [poster]

      Andrew Gambardella, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo

      The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

    • 国際会議
      LLM / NLP

      In-Context Meta Learning Induces Multi-Phase Circuit Emergence

      [paper] [poster] [slide]

      Gouki Minegishi, Hiroki Furuta, Shohei Taniguchi, Yusuke Iwasawa, Yutaka Matsuo

      International Conference on Machine Learning (ICML2025)

    • 国際会議
      LLM / NLP

      Plan-and-Act: Improving Planning of Agents for Long-Horizon Tasks

      [paper]

      Lutfi Eren Erdogan, Nicholas Lee, Sehoon Kim, Suhong Moon, Hiroki Furuta, Gopala Anumanchipalli, Kurt Keutzer, Amir Gholami

      International Conference on Machine Learning (ICML2025)

    • 論文誌
      LLM / NLP

      Continual Pre-training on Character-Level Noisy Texts Makes Decoder-based Language Models Robust Few-shot Learners

      [paper]

      Takeshi Kojima, Yutaka Matsuo, Yusuke Iwasawa

      Transactions of the Association for Computational Linguistics (TACL)

    • 国際会議
      LLM / NLP

      Lost in the Distance: Large Language Models Struggle to Capture Long-Distance Relational Knowledge

      [paper]

      Meiyun Wang, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo

      The 2025 Annual Conference of the Nations of the Americas Chapter of the ACL (NAACL 2025)

    • 国際会議
      LLM / NLP

      Language Models can Categorize System Inputs for Performance Analysis

      [paper]

      Dominic Sobhani, Ruiqi Zhong, Edison Marrese-Taylor, Keisuke Sakaguchi, Yutaka Matsuo

      Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL2025)

    • 国際会議
      LLM / NLP

      Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words

      [paper] [poster]

      Gouki Minegishi, Hiroki Furuta, Yusuke Iwasawa, Yutaka Matsuo

      International Conference on Learning Representations (ICLR 2025)

    • 論文誌
      LLM / NLP

      Exposing Limitations of Language Model Agents in Sequential-Task Compositions on the Web

      [paper] [web]

      Hiroki Furuta, Yutaka Matsuo, Aleksandra Faust, Izzeddin Gur.

      Transactions on Machine Learning Research (TMLR).

    • 受賞
      LLM / NLP

      2024年度 第19回言語処理若手シンポジウム(YANS2024),スポンサー賞(株式会社日立製作所賞): “指示数増加による大規模言語モデルの指示追従性能への悪影響”

      小島 武