HOT RADAR
llmHEAT 7.9°

AI CURATED HOT STORY · 11/17/2026

Prefrontal AI and ETH Zurich Unveil LSA to Speed Ultra-Long Context Attention

1 reports archived1 independent sourcesupdated 11/17/2026, 11:38:00
Synthesis & Latest Updates
1 Sources Cross-Validated

London startup Prefrontal AI and ETH Zurich jointly propose Latent Sparse Attention (LSA), which targets the indexer overhead that dominates sparse attention at extreme context lengths. At 10M tokens it speeds up single-layer prefill by 22.7x on GLM-5.3-Flash and 72.4x on Qwen3.8-27B while keeping accuracy close to the original models, offering a subquadratic path toward 10M-token context.

LATEST/London startup Prefrontal AI and ETH Zurich jointly propose Latent Sparse Attention (LSA), which targets the indexer overhead that dominates sparse attention at extreme context lengths. At 10M tokens it speeds up single-layer prefill by 22.7x on GLM-5.3-Flash and 72.4x on Qwen3.8-27B while keeping accuracy close to the original models, offering a subquadratic path toward 10M-token context.

Hourly heat curve

8 fully observed hours
Now
7.9
Peak
9.422:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first