The Research Desk.

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

Last Brew Time: Jun 14, 2026, 7:28 AM PT

X.com Research Buzz

AI models collapse when trained on recursively generated data
X.com
25206

AI models collapse when trained on recursively generated data

Ilia Shumailov, Zakhar Shumaylov, Yiren Zhao, Yarin Gal, Nicolas Papernot, Ross Anderson

University of Oxford, University of Cambridge, Imperial College London, University of Toronto

NLP
LLMs Reproduce Human Purchase Intent via Semantic Similarity Elicitation of Likert Ratings
X.com
8746

LLMs Reproduce Human Purchase Intent via Semantic Similarity Elicitation of Likert Ratings

Benjamin F. Maier, Ulf Aslak, Luca Fiaschi, Nina Rismal, Kemble Fletcher, Christian C. Luhmann, Robbie Dow, Kli Pappas, Thomas V. Wiecki

PyMC Labs, Colgate-Palmolive Company

AlphaXiv Trending

Self-Harness: Harnesses That Improve Themselves
AlphaXiv
135

Self-Harness: Harnesses That Improve Themselves

Hangfan Zhang, Shao Zhang, Kangcong Li

Shanghai Artificial Intelligence Laboratory

Retrieval
AlphaXiv
82

First Steps Toward Automated AI Research

Recursive Superintelligence

Peking University

From AGI to ASI
AlphaXiv
60

From AGI to ASI

Tim Genewein, Matija Franklin, Alexander Lerchner

Google DeepMind, University of Waterloo (work conducted while at Google DeepMind, Australian National University, University College London, Google DeepMind University of Waterloo (work conducted while at Google DeepMind

Reinforcement Learning
Test-Time Gradient Guidance of Flow Policies in Reinforcement Learning
AlphaXiv
54

Test-Time Gradient Guidance of Flow Policies in Reinforcement Learning

Zhiyuan Zhou, Andy Peng, Charles Xu

MiniMax Sparse Attention
AlphaXiv
51

MiniMax Sparse Attention

Xunhao Lai, Weiqi Xu, Yufeng Yang

Peking University, NVIDIA, Zhejiang University, Huazhong University of Science and Technology, MiniMax Peking University

HuggingFace Daily Papers

NLP
Rethinking Psychometric Evaluation of LLMs: When and Why Self-Reports Predict Behavior
HuggingFace
5

Rethinking Psychometric Evaluation of LLMs: When and Why Self-Reports Predict Behavior

Rafal Kocielnik, Pengrui Han, Peiyang Song, Myrl G. Marmarelis, Ramit Debnath

Reinforcement Learning
See What I See, Know What I Think: Dense Latent Communication Across Heterogeneous Agents
HuggingFace
3

See What I See, Know What I Think: Dense Latent Communication Across Heterogeneous Agents

Siyi Chen, Xiaoyan Zhang, Meng Wu, Jonathan Tremblay, Valts Blukis

University of Michigan

Reinforcement Learning
WebChallenger: A Reliable and Efficient Generalist Web Agent
HuggingFace
2

WebChallenger: A Reliable and Efficient Generalist Web Agent

Jayoo Hwang, Xiaowen Zhang, Vedant Padwal

Reinforcement Learning
Getting Better at Working With You: Compiling User Corrections into Runtime Enforcement for Coding Agents
HuggingFace
2

Getting Better at Working With You: Compiling User Corrections into Runtime Enforcement for Coding Agents

Yujun Zhou, Kehan Guo, Haomin Zhuang, Xiangqi Wang, Yue Huang

NLP
The Cold-Start Safety Gap in LLM Agents
HuggingFace
2

The Cold-Start Safety Gap in LLM Agents

Chung-En Sun, Linbo Liu, Tsui-Wei Weng

NLP
On the Limits of LLM Adaptability: Impact of Model-Internalized Priors on Annotation Task Performance
HuggingFace
0

On the Limits of LLM Adaptability: Impact of Model-Internalized Priors on Annotation Task Performance

Etienne Casanova, Rafal Kocielnik, R. Michael Alvarez

California institute of technology