cd/entity/Xiao et al.· home› entities› Xiao et al.
grep -l @xiao et al. /news/*.json | wc -l → 1

Xiao et al.

mentions 1 type Person feed RSS

// recent coverage 1 mentions

07:55
2026-07-17
dev.to
large-language-models

Attention Sinks: Why Streaming LLMs Break When You Evict Token 0

A developer explains that attention sinks—tokens at position 0 that absorb excess attention weight—cause streaming LLMs to fail when evicted from the KV cache. The softmax normalization forces the mod…

// co-occurs with top 6 entities
StreamingLLM 1 Evan Miller 1 Llama 1 Mistral 1 Qwen 1 GPT 1
// topics top 5 topics
large language models 1 ai research 1 ai infrastructure 1 machine learning 1 neural networks 1