The Research Desk.

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

Last Brew Time: May 28, 2026, 7:43 AM PT

X.com Research Buzz

Generative AI Without Guardrails Can Harm Learning: Evidence from High School Mathematics
X.com
3862

Generative AI Without Guardrails Can Harm Learning: Evidence from High School Mathematics

Hamsa Bastani, Osbert Bastani, Alp Sungu, Haosen Ge, Özge Kabakcı, Rei Mariman

Wharton School, University of Pennsylvania

AlphaXiv Trending

NLP
Do Language Models Need Sleep? Offline Recurrence for Improved Online Inference
AlphaXiv
98

Do Language Models Need Sleep? Offline Recurrence for Improved Online Inference

Sangyun Lee, Sean McLeish, Tom Goldstein

Carnegie Mellon University, University of Maryland

Reinforcement Learning
SkillOpt: Executive Strategy for Self-Evolving Agent Skills
AlphaXiv
96

SkillOpt: Executive Strategy for Self-Evolving Agent Skills

Yifan Yang, Ziyang Gong, Weiquan Huang

Microsoft, Shanghai Jiao Tong University, Tongji University, Fudan University

Speech Audio
Bandwidth-Efficient and Privacy-Preserving Edge-Cloud Many-to-Many Speech Translation
AlphaXiv
45

Bandwidth-Efficient and Privacy-Preserving Edge-Cloud Many-to-Many Speech Translation

Yexing Du, Kaiyuan Liu, Youcheng Pan

This work was supported in part by Harbin Institute of Technology and Pengcheng Laboratory

The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence
AlphaXiv
41

The MiniMax-M2 Series: Mini Activations Unleashing Max Real-World Intelligence

MiniMax, Aili Chen, Aonian Li

When Does LeJEPA Learn a World Model?
AlphaXiv
40

When Does LeJEPA Learn a World Model?

David Klindt, Yann LeCun, Randall Balestriero

Cold Spring Harbor Laboratory, New York University, Brown University

Computer Vision
HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videos
AlphaXiv
40

HumanEgo: Zero-Shot Robot Learning from Minutes of Human Egocentric Videos

Wang, Botao He, Kelin Yu

University of Maryland

HuggingFace Daily Papers

NLP
Rethinking How to Remember: Beyond Atomic Facts in Lifelong LLM Agent Memory
HuggingFace
9

Rethinking How to Remember: Beyond Atomic Facts in Lifelong LLM Agent Memory

Jingwei Sun, Jianing Zhu, Jiangchao Yao, Tongliang Liu, Bo Han

TMLR Group

Machine Learning
GradSentry: Gradient Spectral Entropy for Backdoor Sample Filtering in Large Language Model Fine-Tuning
HuggingFace
8

GradSentry: Gradient Spectral Entropy for Backdoor Sample Filtering in Large Language Model Fine-Tuning

Haodong Zhao, Tianyi Xu, Tianhang Zhao, Zhuosheng Zhang, Gongshen Liu

Shanghai Jiao Tong University

Retrieval
The Fragility of Chain-of-Thought Monitoring Across Typologically Diverse Languages
HuggingFace
4

The Fragility of Chain-of-Thought Monitoring Across Typologically Diverse Languages

Eric Onyame, Runtao Zhou, Kowshik Thopalli, Bhavya Kailkhura, Chirag Agarwal

Aikyam Lab

Compass Group Tue
Models That Know How Evaluations Are Designed Score Safer
HuggingFace
4

Models That Know How Evaluations Are Designed Score Safer

Katharina Deckenbach, Haritz Puerto, Jonas Geiping, Sahar Abdelnabi

COMPASS research group at ELLIS Institute Tübingen

Computer Vision
Everything at Every Scale: Scale-Invariant Diffusion with Continuous Super-Resolution
HuggingFace
4

Everything at Every Scale: Scale-Invariant Diffusion with Continuous Super-Resolution

Zixin Jessie Chen, Zhuo Chen, Archer Wang, Jeff Gore, William T. Freeman

Massachusetts Institute of Technology

NLP
Efficient and Scalable Provenance Tracking for LLM-Generated Code Snippets
HuggingFace
1

Efficient and Scalable Provenance Tracking for LLM-Generated Code Snippets

Andrea Gurioli, Davide D'Ascenzo, Federico Pennino, Maurizio Gabbrielli, Stefano Zacchiroli