The Research Desk.

The most upvoted and starred AI research crossing the community today.

Last Brew Time: Sep 13, 2026, 10:51 AM PT

AlphaXiv Trending

Efficiency
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression
AlphaXiv
268

DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression

NLP
Recurrent Looped Transformer
AlphaXiv
100

Recurrent Looped Transformer

Thinking with Looped Flows
AlphaXiv
57

Thinking with Looped Flows

Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data
AlphaXiv
55

Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data

Affiliation: Stanford University, Affiliation: Paul G. Allen School of Computer Science, University of Washington, Stanford University, Paul G. Allen School of Computer Science, University of Washington

The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement
AlphaXiv
41

The Last AI Built by Humans: Toward Genuine Recursive Self-Improvement

HuggingFace Daily Papers

NLP
Studying Image Tokenizers as Visual Languages in Unified Multimodal Models
HuggingFace
28

Studying Image Tokenizers as Visual Languages in Unified Multimodal Models

Siting Li, Zhengyang Wang, Simon Shaolei Du, Xi Chen, Yang Liu

Amazon Frontier AI & Robotics

Reasoning
Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning
HuggingFace
27

Building Multilingual Bridges: Data Mixing as the Pillar of Generalization for In-Language Reasoning

Mehrnaz Mofakhami, Ananya Sahu, Alejandro R. Salamanca, Daniel D'souza, Alexandre Berard

Cohere Labs

Reinforcement Learning
Beyond Solver Verdicts: Generative Reward Models for Autoformalization
HuggingFace
27

Beyond Solver Verdicts: Generative Reward Models for Autoformalization

Vikash Singh, Debargha Ganguly, Aman Goel, Ali Torkamani, Xiaoxue Han

Amazon Web Services

KaushalrajPuwar
Adaptive Bridge: A Proxy-Based Decoupling Layer for Mitigating DDS Backpressure in ROS 2
HuggingFace
23

Adaptive Bridge: A Proxy-Based Decoupling Layer for Mitigating DDS Backpressure in ROS 2

Kaushalraj Puwar, B. Thangaraju

Machine Learning
ActReview: Rebuttal-Guided Training Data and Rubric Rewards for Actionable Peer Review Generation
HuggingFace
22

ActReview: Rebuttal-Guided Training Data and Rubric Rewards for Actionable Peer Review Generation

Yiling Ma, Yilun Zhao, Sihong Wu, Ziyu Chen, Manasi Patwardhan

Yale NLP Lab

Retrieval
Think Before You Link: Rarity, Reasoning, and Retrieval in Multilingual Entity Linking
HuggingFace
22

Think Before You Link: Rarity, Reasoning, and Retrieval in Multilingual Entity Linking

Parinthapat Pengpun, Simran Khanuja, Graham Neubig

NeuLab @ LTI/CMU