Jaesik Choi

The Grammar of a Polysemantic Neuron: Understanding How Neurons Compress Multiple Concepts

Causal Path Tracing in Transformers

SFdiff : Diffusion Model with Self-Generation for Probabilistic Forecasting

Dummy Consistent Path Method for Robust Attribution with Irrelevant Features

Distributed Knowledge Storage Hypothesis: Evidence from Generalization Failure in Knowledge Editing

Analysis of High-order Interactions in Shapley value for Model Interpretation

Dynamic Banzhaf: Game-Theoretic Attribution with Dynamic Feature-Wise Probabilities

RAP:Reconstruction-Augmented Predictors for Robust Time-series Forecasting

State-Covering Trajectory Stitching for Diffusion Planners

Local Manifold Approximation and Projection for Manifold-Aware Diffusion Planning

Causal Path Tracing in Transformers

Rethinking Shapley Value for Negative Interactions in Non-convex Games

Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning

Preference Banzhaf: A Game-Theoretic Index with Feature-wise Probabilities

TF-score: Time-series Forecasting using score-based diffusion model

Measuring Effects of Steered Representation in Large Language Models