publications

Conference

1. AACL2025

   

   Interpreting Multi-Attribute Confounding through Numerical Attributes in Large Language Models

   Hirohane Takagi^*, Gouki Minegishi^*, Shota Kizawa, Issey Sukeda, and Hitomi Yanaka

   In International Joint Conference on Natural Language Processing & Asia-Pacific Chapter of the Association for Computational Linguistics 2025, 2025

   *Equal contribution

   arXiv
2. Neurips2025

   

   Topology of Reasoning: Understanding Large Reasoning Models through Reasoning Graph Properties

   Gouki Minegishi, Hiroki Furuta, Takeshi Kojima, Yusuke Iwasawa, and Yutaka Matsuo

   In The Thirty-ninth Annual Conference on Neural Information Processing Systems, Neurips, 2025

   arXiv Bib

   @inproceedings{minegishi2025topologyreasoningunderstandinglarge,
     title = {Topology of Reasoning: Understanding Large Reasoning Models through Reasoning Graph Properties},
     author = {Minegishi, Gouki and Furuta, Hiroki and Kojima, Takeshi and Iwasawa, Yusuke and Matsuo, Yutaka},
     booktitle = {The Thirty-ninth Annual Conference on Neural Information Processing Systems, Neurips},
     year = {2025},
     url = {https://openreview.net/forum?id=o1g8NWkxqf},
     kind = {Conference},
   }

3. ICML2025

   

   Beyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit Emergence

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

   In Forty-second International Conference on Machine Learning, ICML, 2025

   arXiv Bib Code

   @inproceedings{anonymous2025incontext,
     title = {Beyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit Emergence},
     author = {Minegishi, Gouki and Furuta, Hiroki and Taniguchi, Shohei and Iwasawa, Yusuke and Matsuo, Yutaka},
     booktitle = {Forty-second International Conference on Machine Learning, ICML},
     year = {2025},
     url = {https://openreview.net/forum?id=Xw01vF13aV},
     kind = {Conference}
   }

4. ICLR2025

   

   Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words

   Gouki Minegishi, Hiroki Furuta, Yusuke Iwasawa, and Yutaka Matsuo

   In The Thirteenth International Conference on Learning Representations, ICLR, 2025

   arXiv Bib Code

   @inproceedings{minegishi2025rethinking,
     title = {Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words},
     author = {Minegishi, Gouki and Furuta, Hiroki and Iwasawa, Yusuke and Matsuo, Yutaka},
     booktitle = {The Thirteenth International Conference on Learning Representations, ICLR},
     year = {2025},
     url = {https://openreview.net/forum?id=HpUs2EXjOl},
     kind = {Conference}
   }

5. Neurips2024

   ADOPT: Modified Adam Can Converge with Any {}beta_2 with the Optimal Rate

   Shohei Taniguchi, Keno Harada, Gouki Minegishi, Yuta Oshima, Seong Cheol Jeong, and 5 more authors

   In The Thirty-eighth Annual Conference on Neural Information Processing Systems, Neurips, 2024

   Bib

   @inproceedings{taniguchi2024adopt,
     title = {{ADOPT}: Modified Adam Can Converge with Any \${\textbackslash}beta\_2\$ with the Optimal Rate},
     author = {Taniguchi, Shohei and Harada, Keno and Minegishi, Gouki and Oshima, Yuta and Jeong, Seong Cheol and Nagahara, Go and Iiyama, Tomoshi and Suzuki, Masahiro and Iwasawa, Yusuke and Matsuo, Yutaka},
     booktitle = {The Thirty-eighth Annual Conference on Neural Information Processing Systems, Neurips},
     year = {2024},
     url = {https://openreview.net/forum?id=rzvVm0LsyK},
     kind = {Conference}
   }

Conference Workshop

1. ICLR2025 WS

   In-Context Meta Learning Induces Multi-Phase Circuit Emergence

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

   In ICLR 2025 Workshop on Building Trust in Language Models and Applications, 2025

   Awarded arXiv Bib

   Oral Presentations

   @inproceedings{minegishi2025incontext,
     title = {In-Context Meta Learning Induces Multi-Phase Circuit Emergence},
     author = {Minegishi, Gouki and Furuta, Hiroki and Taniguchi, Shohei and Iwasawa, Yusuke and Matsuo, Yutaka},
     booktitle = {ICLR 2025 Workshop on Building Trust in Language Models and Applications},
     year = {2025},
     url = {https://openreview.net/forum?id=LNMfzv8TNb},
     kind = {Conference Workshop}
   }

2. ICLR2024 WS

   Interpreting Grokked Transformers in Complex Modular Arithmetic

   Hiroki Furuta, Minegishi Gouki, Yusuke Iwasawa, and Yutaka Matsuo

   In ICLR 2024 Workshop on Bridging the Gap Between Practice and Theory in Deep Learning, 2024

   Awarded arXiv Bib Code

   Oral Presentations

   @inproceedings{furuta2024interpreting,
     title = {Interpreting Grokked Transformers in Complex Modular Arithmetic},
     author = {Furuta, Hiroki and Gouki, Minegishi and Iwasawa, Yusuke and Matsuo, Yutaka},
     booktitle = {ICLR 2024 Workshop on Bridging the Gap Between Practice and Theory in Deep Learning},
     year = {2024},
     url = {https://openreview.net/forum?id=bP8qHYBoTk},
     kind = {Conference Workshop}
   }

3. ICLR2024 WS

   Bridging Lottery ticket and Grokking: Is Weight Norm Sufficient to Explain Delayed Generalization?

   Minegishi Gouki, Yusuke Iwasawa, and Yutaka Matsuo

   In ICLR 2024 Workshop on Bridging the Gap Between Practice and Theory in Deep Learning, 2024

   Bib

   @inproceedings{gouki2024bridging,
     title = {Bridging Lottery ticket and Grokking: Is Weight Norm Sufficient to Explain Delayed Generalization?},
     author = {Gouki, Minegishi and Iwasawa, Yusuke and Matsuo, Yutaka},
     booktitle = {ICLR 2024 Workshop on Bridging the Gap Between Practice and Theory in Deep Learning},
     year = {2024},
     url = {https://openreview.net/forum?id=YvpL42Bfh5},
     kind = {Conference Workshop}
   }

Journal

1. TMLR2025

   

   Bridging Lottery Ticket and Grokking: Understanding Grokking from Inner Structure of Networks

   Gouki Minegishi, Yusuke Iwasawa, and Yutaka Matsuo

   Transactions on Machine Learning Research, TMLR, 2025

   arXiv Bib Code

   @article{minegishi2025bridging,
     title = {Bridging Lottery Ticket and Grokking: Understanding Grokking from Inner Structure of Networks},
     author = {Minegishi, Gouki and Iwasawa, Yusuke and Matsuo, Yutaka},
     journal = {Transactions on Machine Learning Research, TMLR},
     issn = {2835-8856},
     year = {2025},
     url = {https://openreview.net/forum?id=eQeYyup1tm},
     note = {},
     kind = {Journal}
   }

2. TMLR2024

   Towards Empirical Interpretation of Internal Circuits and Properties in Grokked Transformers on Modular Polynomials

   Hiroki Furuta, Gouki Minegishi, Yusuke Iwasawa, and Yutaka Matsuo

   Transactions on Machine Learning Research, TMLR, 2024

   arXiv Bib

   @article{furuta2024towards,
     title = {Towards Empirical Interpretation of Internal Circuits and Properties in Grokked Transformers on Modular Polynomials},
     author = {Furuta, Hiroki and Minegishi, Gouki and Iwasawa, Yusuke and Matsuo, Yutaka},
     journal = {Transactions on Machine Learning Research, TMLR},
     issn = {2835-8856},
     year = {2024},
     url = {https://openreview.net/forum?id=MzSf70uXJO},
     note = {},
     kind = {Journal}
   }

Preprint

1. Preprint

   

   Mechanism of Task-oriented Information Removal in In-context Learning

   Hakaze Cho, Haolin Yang, Gouki Minegishi, and Naoya Inoue

   2025

   arXiv
2. Preprint

   

   RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMs

   Kohsei Matsutani, Shota Takashiro, Gouki Minegishi, Takeshi Kojima, Yusuke Iwasawa, and 1 more author

   2025

   arXiv