StartLux

company

https://x.com/StartLuxAI

AI & ML interests

None defined yet.

Recent Activity

zunhai authored a paper 1 day ago

QuantMLA: Function-Aligned Dual-Path Quantization for Low-Bit MLA KV Caching

zunhai submitted a paper 1 day ago

QuantMLA: Function-Aligned Dual-Path Quantization for Low-Bit MLA KV Caching

Lie24 authored a paper 2 months ago

Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus

View all activity

Papers

Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus

View all Papers

startlux-models 's papers 1

Submitted by

zunhaisu

Massive Activations in Hybrid Linear Attention Large Language Models: Pre-Attention Spikes and Inter-Spike Plateaus

startlux-models