Author

Youn-Sik Hong

2 papers indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#artificial intelligence Preprint Aug 2026

The Mask Is Not the Model: Auditing Prefix Invariance in Attention, State-Space, and Hybrid Sequence Models

A lightweight audit is given of attention-mask inspection, two forward passes, no training or gradients, yielding a per-layer score localizing where causality breaks, yielding a per-layer score localizing where causality breaks.

Taebong Kim, Youn-Sik Hong, Minsik Kim et al. · 1 citation

#machine learning Preprint Jul 2026

Placement Is Free, Composition Is Not: The Latin Square as a Provably-Balanced Construction for Heterogeneous Sequence-Mixer Stacks

The introduction of Aether-7B-5Attn, a 6.59B-parameter mixture-of-experts model whose 49 layers contain seven sequence-mixing mechanisms arranged as a Latin square, indicates that performance depends primarily on heterogeneous composition distributed across depth rather than on any particular permutation.

Taebong Kim, Youn-Sik Hong, Minsik Kim et al. · 0 citations