Author

Jian Li

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.

#machine learning Jun 2026

On the Residual Scaling of Looped Transformers: Stability and Transferability

A key consequence is that the optimal learning rate depends only on the number of unique layers $L$, not on the loop count $N, enabling direct hyperparameter transfer from small to large $N$ without retuning.

Shaowen Wang, Bing-Rui Li, Ge Zhang et al. · arXiv.org · 7 citations · ⚡1

#machine learning Preprint Sep 2026

SMELT: Scaling Laws for Compute-Matched MoE Looped Transformers

Results show that looping can improve Transformers even under budget matching, offering a practical recipe that turns depth reuse into measurable gains.

Shaowen Wang, Ge Zhang, Kai-Rong Luo et al. · 17 citations