LLM-Drop

AI & ML interests

Efficient and adaptive foundation models across language and multimodal intelligence.

Recent Activity

shwai-he  authored a paper about 5 hours ago

Selective Reflection-Tuning: Student-Selected Data Recycling for LLM Instruction-Tuning

shwai-he  authored a paper about 5 hours ago

What Matters in Transformers? Not All Attention is Needed

shwai-he  authored a paper about 5 hours ago

Capacity-Aware Inference: Mitigating the Straggler Effect in Mixture of Experts

View all activity

Papers

Demystifying When Pruning Works via Representation Hierarchies

Understanding and Harnessing Sparsity in Unified Multimodal Models

View all Papers

Team members 2