Gouki Minegishi

19 papers A* 3Journal 15Unranked 1
YearRankTypeTitle / Venue / Authors
2026 J jnl
Emergent Analogical Reasoning in Transformers.
CoRR
Gouki Minegishi, Jingyuan Feng, Hiroki Furuta, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo
2026 J jnl
LIT-RAGBench: Benchmarking Generator Capabilities of Large Language Models in Retrieval-Augmented Generation.
CoRR
Koki Itai, Shunichi Hasegawa, Yuta Yamamoto, Gouki Minegishi, Masaki Otsuki
2026 J jnl
Safe Transformer: An Explicit Safety Bit For Interpretable And Controllable Alignment.
CoRR
Jingyuan Feng, Andrew Gambardella, Gouki Minegishi, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo
2026 J jnl
Steering at the Source: Style Modulation Heads for Robust Persona Control.
CoRR
Yoshihiro Izawa, Gouki Minegishi, Koshi Eguchi, Sosuke Hosokawa, Kenjiro Taura
2025 A* conf
Beyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit Emergence.
ICML
Gouki Minegishi, Hiroki Furuta, Shohei Taniguchi, Yusuke Iwasawa, Yutaka Matsuo
2025 J jnl
Beyond Induction Heads: In-Context Meta Learning Induces Multi-Phase Circuit Emergence.
CoRR
Gouki Minegishi, Hiroki Furuta, Shohei Taniguchi, Yusuke Iwasawa, Yutaka Matsuo
2025 J jnl
Bridging Lottery Ticket and Grokking: Understanding Grokking from Inner Structure of Networks.
Trans. Mach. Learn. Res.
Gouki Minegishi, Yusuke Iwasawa, Yutaka Matsuo
2025 — conf
Interpreting Multi-Attribute Confounding through Numerical Attributes in Large Language Models.
IJCNLP-AACL (long papers)
Hirohane Takagi, Gouki Minegishi, Shota Kizawa, Issey Sukeda, Hitomi Yanaka
2025 J jnl
Interpreting Multi-Attribute Confounding through Numerical Attributes in Large Language Models.
CoRR
Hirohane Takagi, Gouki Minegishi, Shota Kizawa, Issey Sukeda, Hitomi Yanaka
2025 J jnl
Mechanism of Task-oriented Information Removal in In-context Learning.
CoRR
Hakaze Cho, Haolin Yang, Gouki Minegishi, Naoya Inoue
2025 J jnl
RL Squeezes, SFT Expands: A Comparative Study of Reasoning LLMs.
CoRR
Kohsei Matsutani, Shota Takashiro, Gouki Minegishi, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo
2025 A* conf
Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words.
ICLR
Gouki Minegishi, Hiroki Furuta, Yusuke Iwasawa, Yutaka Matsuo
2025 J jnl
Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words.
CoRR
Gouki Minegishi, Hiroki Furuta, Yusuke Iwasawa, Yutaka Matsuo
2025 J jnl
Topology of Reasoning: Understanding Large Reasoning Models through Reasoning Graph Properties.
CoRR
Gouki Minegishi, Hiroki Furuta, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo
2024 A* conf
ADOPT: Modified Adam Can Converge with Any β
NeurIPS
Shohei Taniguchi, Keno Harada, Gouki Minegishi, Yuta Oshima, Seong Cheol Jeong, Go Nagahara, Tomoshi Iiyama, Masahiro Suzuki, Yusuke Iwasawa, Yutaka Matsuo
2024 J jnl
ADOPT: Modified Adam Can Converge with Any β
CoRR
Shohei Taniguchi, Keno Harada, Gouki Minegishi, Yuta Oshima, Seong Cheol Jeong, Go Nagahara, Tomoshi Iiyama, Masahiro Suzuki, Yusuke Iwasawa, Yutaka Matsuo
2024 J jnl
Interpreting Grokked Transformers in Complex Modular Arithmetic.
CoRR
Hiroki Furuta, Gouki Minegishi, Yusuke Iwasawa, Yutaka Matsuo
2024 J jnl
Towards Empirical Interpretation of Internal Circuits and Properties in Grokked Transformers on Modular Polynomials.
Trans. Mach. Learn. Res.
Hiroki Furuta, Gouki Minegishi, Yusuke Iwasawa, Yutaka Matsuo
2023 J jnl
Grokking Tickets: Lottery Tickets Accelerate Grokking.
CoRR
Gouki Minegishi, Yusuke Iwasawa, Yutaka Matsuo