Daiki Chijiwa

Lossless Vocabulary Reduction for Auto-Regressive Language Models

Do We Really Need Permutations? Impact of Model Width on Linear Mode Connectivity

Zero-shot Concept Bottleneck Models

Plausible Token Amplification for Improving Accuracy of Differentially Private In-Context Learning Based on Implicit Bayesian Inference

Portable Reward Tuning: Towards Reusable Fine-Tuning across Different Pretrained Models

Zero-shot Concept Bottleneck Models via Sparse Regression of Retrieved Concepts