Mechanistic - a yicui Collection

yicui 's Collections

Coding

ICL

RL

TDD

Theory

Mechanistic

updated about 17 hours ago

Massive Activations in Large Language Models

Paper • 2402.17762 • Published Feb 27 • 1
What Matters in Transformers? Not All Attention is Needed

Paper • 2406.15786 • Published Jun 22 • 29
The Super Weight in Large Language Models

Paper • 2411.07191 • Published 11 days ago • 2
Top-nσ: Not All Logits Are You Need

Paper • 2411.07641 • Published 10 days ago • 15