·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
MAGNET: Manifold-Aware Graph Diffusion Network for Connectome Generation33m◆Counterfactual Model Testing Under Distribution Shift33m◆Centralized Critics for N-Agent Ad Hoc Teamwork33m◆Where and What to Reward? Understanding Generalization Between Verifiable and Open-Ended Domains in Reinforcement Learning33m◆Fast KVzip: Efficient and Accurate LLM Inference with Gated KV Eviction33m◆Numina Fuse: Interactive Autoformalization in Lean 433m◆On the Limits of Sparse Linear Model with Shuffled Differential Privacy33m◆Video World Models Need Predictable Latents33m◆AutoPRL: Automated Path Representation Learning via a Learnable View Generator33m◆Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models33m◆Latent Goal Prediction from Language for Model-Based Planning33m◆Rectified Language Flow33m◆Trainable Boundary-Keeping Where No Formal Answer Exists: A Controlled Knock-In Study of Reward Training in Evidence-Sufficiency Judgment33m◆Intrinsic Goals for Autonomous Agents: Hierarchical Exploration Captures Future Volitional Action in Drosophila33m◆Estimating Long-term Heterogeneous Dose-response Curve: Generalization Bound Leveraging Optimal Transport Weights33m◆Intervention-Certified Amplification: A Validity Gate for $\sigma$-Weighted Post-Training33m◆Single-Cycle Battery Remaining Useful Life Prediction with Compressed Continuous Wavelet Features and Deep Neural Networks33m◆DoubleFree:Exploiting Exploration History for Rollout-Free Multi-Task LLMs Training33m◆Unbiased Alignment for Large Language Models with Noisy Preferences33m◆ExCal: Optimization-Free Multivariate Time Series Anomaly Detection via Excess Calibration33m◆MAGNET: Manifold-Aware Graph Diffusion Network for Connectome Generation33m◆Counterfactual Model Testing Under Distribution Shift33m◆Centralized Critics for N-Agent Ad Hoc Teamwork33m◆Where and What to Reward? Understanding Generalization Between Verifiable and Open-Ended Domains in Reinforcement Learning33m◆Fast KVzip: Efficient and Accurate LLM Inference with Gated KV Eviction33m◆Numina Fuse: Interactive Autoformalization in Lean 433m◆On the Limits of Sparse Linear Model with Shuffled Differential Privacy33m◆Video World Models Need Predictable Latents33m◆AutoPRL: Automated Path Representation Learning via a Learnable View Generator33m◆Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models33m◆Latent Goal Prediction from Language for Model-Based Planning33m◆Rectified Language Flow33m◆Trainable Boundary-Keeping Where No Formal Answer Exists: A Controlled Knock-In Study of Reward Training in Evidence-Sufficiency Judgment33m◆Intrinsic Goals for Autonomous Agents: Hierarchical Exploration Captures Future Volitional Action in Drosophila33m◆Estimating Long-term Heterogeneous Dose-response Curve: Generalization Bound Leveraging Optimal Transport Weights33m◆Intervention-Certified Amplification: A Validity Gate for $\sigma$-Weighted Post-Training33m◆Single-Cycle Battery Remaining Useful Life Prediction with Compressed Continuous Wavelet Features and Deep Neural Networks33m◆DoubleFree:Exploiting Exploration History for Rollout-Free Multi-Task LLMs Training33m◆Unbiased Alignment for Large Language Models with Noisy Preferences33m◆ExCal: Optimization-Free Multivariate Time Series Anomaly Detection via Excess Calibration33m◆
News/HyScale: Scaling Long-Context Training for Learned Sparse Attention
arxiv
PublishedOctober 26, 2026 at 4:00 AM

HyScale: Scaling Long-Context Training for Learned Sparse Attention

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2610.32057v1 Announce Type: new Abstract: We present HyScale, a long-context training system for learned sparse attention. Learned sparse attention is a promising architecture for million-token contexts, widely adopted in production since DeepSeek-V3.2. Selection, however, grows into the train

Stay posted· Newsletter

Three short emails a week: Monday, Wednesday and Friday. Top movers, new models and the most-referenced story.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivMAGNET: Manifold-Aware Graph Diffusion Network for Connectome Generation33marxivCounterfactual Model Testing Under Distribution Shift33marxivCentralized Critics for N-Agent Ad Hoc Teamwork33marxivWhere and What to Reward? Understanding Generalization Between Verifiable and Open-Ended Domains in Reinforcement Learning33m
The Bubble Brief
MON · WED · FRI

Read AI insights every Monday, Wednesday and Friday — top movers, new models, story of the week.

// no spam · unsubscribe one-click · free forever

Your email is used only to send The Bubble Brief. Privacy policy

Originally published on arxiv ↗
PrivacyTermsRefundsLegal notice|AboutMethodologyContact

© 2026 Databubble

Built by Marouane Gazouzi · Readyo, AI interview coach

HomeModelsNews