Latent Sparse Attention: Towards Subquadratic Ultra-Long Context Attention

3 days agoReported by arXiv22

Latent Sparse Attention: Towards Subquadratic Ultra-Long Context Attention. The story centres on Qwen. Reported by arXiv. Bharat Hunt files it under AI Models and AI Agents — the section covering a new or updated model, its capabilities, benchmarks or availability.

Written by Bharat Hunt from the headline and the coverage below. The original reporting is the source of truth.

Read original coverage

Companies, models and people in this story

Qwenmodel

Reported by arXiv

First seen 3 days ago

arXivLatent Sparse Attention: Towards Subquadratic Ultra-Long Context Attention3 days agoAles Kubicek

AI Models23

Selective Transfer of RL Updates for Visual Reasoning

Selective Transfer of RL Updates for Visual Reasoning. It centres on Benchmarks, and also names Qwen. Reported by arXiv. Bharat Hunt files it under AI Models and Computer Vision — the section covering a new or updated model, its capabilities, benchmarks or availability.

arXiv6 Oct 2026

Robotics23

LLMs are General Asynchronous Agents

LLMs are General Asynchronous Agents. It centres on Inference, and also names Qwen. Reported by arXiv. Bharat Hunt files it under Robotics and AI Tools — the section covering embodied AI, humanoids, autonomous vehicles and drones. Open the original report for the full detail.

← All AI stories

Latent Sparse Attention: Towards Subquadratic Ultra-Long Context Attention

Companies, models and people in this story

Reported by arXiv

Selective Transfer of RL Updates for Visual Reasoning

LLMs are General Asynchronous Agents

Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiency

Stale-Document Poisoning: When Outdated Retrieval Overrides Correct Model Answers

Companies, models and people in this story

Reported by arXiv

Related AI coverage

Selective Transfer of RL Updates for Visual Reasoning

LLMs are General Asynchronous Agents

Learning to Stop without Learning to Stop: Self-Supervised Confidence Training Improves Reasoning Efficiency

Stale-Document Poisoning: When Outdated Retrieval Overrides Correct Model Answers