RangeSWA: Training-Free Length Extrapolation for Continuous Diffusion Language Models

2 days agoReported by arXiv24

RangeSWA: Training-Free Length Extrapolation for Continuous Diffusion Language Models. The story centres on Perplexity. Reported by arXiv. Bharat Hunt files it under Generative AI and AI Models — the section covering text, image, audio and video generation — products and their output.

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

Perplexitycompany

Reported by arXiv

First seen 2 days ago

arXivRangeSWA: Training-Free Length Extrapolation for Continuous Diffusion Language Models2 days agoGuangfu Guo

AI Models24

MS-GLA: Multi-Scale Gated Linear Attention for Addressing Representational Bottlenecks via Multi-Temporal Resolution

MS-GLA: Multi-Scale Gated Linear Attention for Addressing Representational Bottlenecks via Multi-Temporal Resolution. It centres on Perplexity, and also names Benchmarks. Reported by arXiv. Bharat Hunt files it under AI Models — the section covering a new or updated model, its capabilities, benchmarks or availability.

arXiv28 Sept 2026

Generative AI23

Representation-Space MMD for Diffusion Language Models

Representation-Space MMD for Diffusion Language Models. It centres on Perplexity, and also names Benchmarks. Reported by arXiv. Bharat Hunt files it under Generative AI and AI Models — the section covering text, image, audio and video generation — products and their output.

← All AI stories

RangeSWA: Training-Free Length Extrapolation for Continuous Diffusion Language Models

Companies, models and people in this story

Reported by arXiv

MS-GLA: Multi-Scale Gated Linear Attention for Addressing Representational Bottlenecks via Multi-Temporal Resolution

Representation-Space MMD for Diffusion Language Models

Differentially Private Mixing of Public Datasets Improves Private Learning

SOL: Measuring Gaps between Text Distributions by Double Sliced Wasserstein Metrics

Companies, models and people in this story

Reported by arXiv

Related AI coverage

MS-GLA: Multi-Scale Gated Linear Attention for Addressing Representational Bottlenecks via Multi-Temporal Resolution

Representation-Space MMD for Diffusion Language Models

Differentially Private Mixing of Public Datasets Improves Private Learning

SOL: Measuring Gaps between Text Distributions by Double Sliced Wasserstein Metrics