← Search

Jishan Hu

2 accepted papers

2025

GPAS: Accelerating Convergence of LLM Pretraining via Gradient-Preserving Activation Scaling

NeurIPS 2025poster

Modern Large Language Models, such as the LLaMA, Qwen and DeepSeek series, predominantly adopt the Pre-LayerNorm (Pre-LN) Transformer architecture. While being stable during pretraining and scalable to large model sizes, Pre-LN suffers from an exponential growth in activation variance across layers,…

Cited by 0SourcecodeScholar
2025

UGMathBench: A Diverse and Dynamic Benchmark for Undergraduate-Level Mathematical Reasoning with Large Language Models

ICLR 2025poster

Large Language Models (LLMs) have made significant strides in mathematical reasoning, underscoring the need for a comprehensive and fair evaluation of their capabilities. However, existing benchmarks often fall short, either lacking extensive coverage of undergraduate-level mathematical problems or…

Cited by 1SourcePDFScholar