The Research Desk.

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

Last Updated: Apr 6, 2026, 12:00 PM PT

X.com Research Buzz

Sycophantic Chatbots Cause Delusional Spiraling, Even in Ideal Bayesians
X.com
3110

Sycophantic Chatbots Cause Delusional Spiraling, Even in Ideal Bayesians

Kartik Chandra, Max Kleiman-Weiner, Jonathan Ragan-Kelley +1 more

The Microarchitecture of DOJO, Tesla's Exa-Scale Computer
X.com
2431

The Microarchitecture of DOJO, Tesla's Exa-Scale Computer

E. Talpes, D. D. Sarma, Anant Nivarti

AlphaXiv Trending

Meta-Harness: End-to-End Optimization of Model Harnesses
AlphaXiv
283

Meta-Harness: End-to-End Optimization of Model Harnesses

Yoonho Lee, Roshen Nair, Qizheng Zhang

#efficiency#alphaxiv
Embarrassingly Simple Self-Distillation Improves Code Generation
AlphaXiv
116

Embarrassingly Simple Self-Distillation Improves Code Generation

Ruixiang Zhang, Richard He Bai, Huangjie Zheng

#efficiency#reasoning#alphaxiv
The Latent Space: Foundation, Evolution, Mechanism, Ability, and Outlook
AlphaXiv
88

The Latent Space: Foundation, Evolution, Mechanism, Ability, and Outlook

Xinlei Yu, Zhangquan Chen, Yongbo He

#alphaxiv
Screening Is Enough
AlphaXiv
87

Screening Is Enough

Ken M. Nakanishi

#alphaxiv
SKILL0: In-Context Agentic Reinforcement Learning for Skill Internalization
AlphaXiv
71

SKILL0: In-Context Agentic Reinforcement Learning for Skill Internalization

Zhengxi Lu, Zhiyuan Yao, Jinyang Wu

#reinforcement-learning#alphaxiv
A Survey of On-Policy Distillation for Large Language Models
AlphaXiv
43

A Survey of On-Policy Distillation for Large Language Models

Mingyang Song, Mao Zheng

#nlp#reinforcement-learning#efficiency#alphaxiv

HuggingFace Daily Papers

Test-Time Scaling Makes Overtraining Compute-Optimal
HuggingFace
12

Test-Time Scaling Makes Overtraining Compute-Optimal

Nicholas Roberts, Sungjun Cho, Zhiqi Gao +2 more

#machine-learning
Communicating about Space: Language-Mediated Spatial Integration Across Partial Views
HuggingFace
10

Communicating about Space: Language-Mediated Spatial Integration Across Partial Views

Ankur Sikarwar, Debangan Mishra, Sudarshan Nikhil +2 more

#ankursikarwar
Swift-SVD: Theoretical Optimality Meets Practical Efficiency in Low-Rank LLM Compression
HuggingFace
3

Swift-SVD: Theoretical Optimality Meets Practical Efficiency in Low-Rank LLM Compression

Ruoling Qi, Yirui Liu, Xuaner Wu +2 more

#nlp#efficiency
VLMs Need Words: Vision Language Models Ignore Visual Detail In Favor of Semantic Anchors
HuggingFace
1

VLMs Need Words: Vision Language Models Ignore Visual Detail In Favor of Semantic Anchors

Haz Sameen Shahgir, Xiaofu Chen, Yu Fu +2 more

#nlp#computer-vision#multimodal
Salt: Self-Consistent Distribution Matching with Cache-Aware Training for Fast Video Generation
HuggingFace
1

Salt: Self-Consistent Distribution Matching with Cache-Aware Training for Fast Video Generation

Xingtong Ge, Yi Zhang, Yushi Huang +2 more

#machine-learning#computer-vision#XingtongGe
Do World Action Models Generalize Better than VLAs? A Robustness Study
HuggingFace
1

Do World Action Models Generalize Better than VLAs? A Robustness Study

Zhanguang Zhang, Zhiyuan Li, Behnam Rahmati +2 more