9 papers
ID Balancing: Stable Training of Extremely Sparse MoE via PID-Based Load Control
Peng Jin, Zihan Qiu, Zekun Wang +8
Scaling Large Language Models (LLMs) via Mixture-of-Experts (MoE) enables massive parameter growth with nearly constant per-token computation. However, further scaling the paramete…
On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability
Zihan Qiu, Zekun Wang, Xiao Li +33
We describe the architecture and ablations of Qwen3.8-Flash-Next, a sparse mixture-of-experts model with 125B parameters, 6B activated per token, and additional 51B parameters of n…
A Unified View of Attention and Residual Sinks: Outlier-Driven Rescaling is Essential for Transformer Training
Zihan Qiu, Zeyu Huang, Kaiyue Wen +16
We investigate the functional role of emergent outliers in large language models, specifically attention sinks (a few tokens that consistently receive large attention logits) and r…
UniAttn: Reducing Inference Costs via Softmax Unification for Post-Training LLMs
Yizhe Xiong, Wei Huang, Xin Ye +6
Post-training is essential for adapting Large Language Models (LLMs) to real-world applications. Deploying post-trained models faces significant challenges due to substantial memor…
Breaking the Stage Barrier: A Novel Single-Stage Approach to Long Context Extension for Large Language Models
Haoran Lian, Junmin Chen, Wei Huang +8
Recently, Large language models (LLMs) have revolutionized Natural Language Processing (NLP). Pretrained LLMs, due to limited training context size, struggle with handling long tok…
LBPE: Long-token-first Tokenization to Improve Large Language Models
Haoran Lian, Yizhe Xiong, Zijia Lin +5
The prevalent use of Byte Pair Encoding (BPE) in Large Language Models (LLMs) facilitates robust handling of subword units and avoids issues of out-of-vocabulary words. Despite its…
MaskMoE: Boosting Token-Level Learning via Routing Mask in Mixture-of-Experts
Zhenpeng Su, Zijia Lin, Xue Bai +8
Scaling the size of a model enhances its capabilities but significantly increases computation complexity. Mixture-of-Experts models (MoE) address the issue by allowing model size t…
Scaffold-BPE: Enhancing Byte Pair Encoding for Large Language Models with Simple and Effective Scaffold Token Removal
Haoran Lian, Yizhe Xiong, Jianwei Niu +7
Byte Pair Encoding (BPE) serves as a foundation method for text tokenization in the Natural Language Processing (NLP) field. Despite its wide adoption, the original BPE algorithm h…
Temporal Scaling Law for Large Language Models
Yizhe Xiong, Xiansheng Chen, Xin Ye +8
Recently, Large Language Models (LLMs) have been widely adopted in a wide range of tasks, leading to increasing attention towards the research on how scaling LLMs affects their per…