collaborators

5 papers

cs.LG2026

The Curse and Blessing of Mean Bias in FP4-Quantized LLM Training

Hengjie Cao, Zhendong Huang, Mengyi Chen +15

FP4 training promises substantial memory and compute savings for large language models, but remains fragile because blockwise quantization is dictated by extreme activation magnitu…

cs.AI2026

DecodeShare: Tracing the Shared Subspace of LLM Decode-Time Decisions

Zishan Shao, Lixun Zhang, Kangning Cui +10

Large language models (LLMs) handle many tasks with one set of parameters, but under KV-cached inference it is unclear what task-general structure, if any, is used at decode time r…

cs.CR2026

Multi-Stream Perturbation Attack: Breaking Safety Alignment of Thinking LLMs Through Concurrent Task Interference

Fan Yang

The widespread adoption of thinking mode in large language models (LLMs) has significantly enhanced complex task processing capabilities while introducing new security risks. When…

cs.LG2026

Spectra: Rethinking Optimizers for LLMs Under Spectral Anisotropy

Zhendong Huang, Hengjie Cao, Fang Dong +14

Gradient signals in LLM training are highly anisotropic: recurrent linguistic structure concentrates energy into a small set of dominant spectral directions, while context specific…

cs.LG2025

Metis: Training LLMs with FP4 Quantization

Hengjie Cao, Mengyi Chen, Yifeng Yang +13

This work identifies anisotropy in the singular value spectra of parameters, activations, and gradients as the fundamental barrier to low-bit training of large language models (LLM…