Hidenori Tanaka

14

Papers

157

Total Citations

Papers (14)

Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks

ICLR: In-Context Learning of Representations

A Percolation Model of Emergence: Analyzing Transformers Trained on a Formal Language

Forking Paths in Neural Text Generation

In-Context Learning Strategies Emerge Rationally

NeurIPS 2025arXiv

Competition Dynamics Shape Algorithmic Phases of In-Context Learning

Swing-by Dynamics in Concept Learning and Compositional Generalization

Towards an Understanding of Stepwise Inference in Transformers: A Synthetic Graph Navigation Model

Compositional Capabilities of Autoregressive Transformers: A Study on Synthetic, Interpretable Tasks

Pruning neural networks without any data by iteratively conserving synaptic flow

NeurIPS 2020arXiv

Beyond BatchNorm: Towards a Unified Understanding of Normalization in Deep Learning

NeurIPS 2021arXiv

Noether’s Learning Dynamics: Role of Symmetry Breaking in Neural Networks

NeurIPS 2021arXiv

Compositional Abilities Emerge Multiplicatively: Exploring Diffusion Models on a Synthetic Task

NeurIPS 2023arXiv

CORNN: Convex optimization of recurrent neural networks for rapid inference of neural dynamics

NeurIPS 2023arXiv