The Research Desk.

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

Last Brew Time: Oct 27, 2026, 10:59 AM PT

X.com Research Buzz

NLP
Why are all LLMs Obsessed with Japanese Culture? On the Hidden Cultural and Regional Biases of LLMs
X.com
12692

Why are all LLMs Obsessed with Japanese Culture? On the Hidden Cultural and Regional Biases of LLMs

NLP
The Algorithmic Self-Portrait: Deconstructing Memory in ChatGPT
X.com
7286

The Algorithmic Self-Portrait: Deconstructing Memory in ChatGPT

Evidence for the Representation of Non-Hierarchical Structures in Language
X.com
7272

Evidence for the Representation of Non-Hierarchical Structures in Language

Morten H. Christiansen, Yngwie A. Nielsen

Cornell University, Aarhus University

AlphaXiv Trending

The Exponential Time Hypothesis Is False
AlphaXiv
96

The Exponential Time Hypothesis Is False

7k, Views

Reinforcement Learning
Memory Tax: Dissecting the Net Value of Agent Memory
AlphaXiv
35

Memory Tax: Dissecting the Net Value of Agent Memory

Views

NLP
How High Can GPT-6 Astra Climb in Contact-Rich Manipulation? PhysLadder: A Hierarchical, Matched Sim-Real Benchmark for Physical Interaction
AlphaXiv
16

How High Can GPT-6 Astra Climb in Contact-Rich Manipulation? PhysLadder: A Hierarchical, Matched Sim-Real Benchmark for Physical Interaction

Views

The University of Hong Kong, Chinese Academy of Sciences, Nanyang Technological University, University of Michigan, Ann Arbor, Southeast University

Reinforcement Learning
Work Agents: Systems, Capabilities, and Optimization
AlphaXiv
15

Work Agents: Systems, Capabilities, and Optimization

Views

Reinforcement Learning
Profiling a Frontier Agent with EBench: From Direct Control to Harness with Policy
AlphaXiv
11

Profiling a Frontier Agent with EBench: From Direct Control to Harness with Policy

Shanghai Artificial Intelligence Laboratory Shanghai Jiao Tong University, Fudan University Zhejiang University

Robotics
Learning Force-Reactive Humanoid Loco-Manipulation
AlphaXiv
7

Learning Force-Reactive Humanoid Loco-Manipulation

Affiliation: Amazon FAR (Frontier AI & Robotics) New York University Carnegie Mellon University University of Chicago, Amazon FAR team co-lead, Amazon FAR (Frontier AI & Robotics) New York University Carnegie Mellon University University of Chicago

HuggingFace Daily Papers

Reinforcement Learning
CRM-Agent-Bench: Benchmarking Tool-Using Agents in Stateful Enterprise Environments
HuggingFace
3

CRM-Agent-Bench: Benchmarking Tool-Using Agents in Stateful Enterprise Environments

Monojit Banerjee, Zhujun Lan, Defu Cao

Salesforce AI Cloud

Reinforcement Learning
iPrefBench: A User-Centric Benchmark for Personalized Mobile Use Agents
HuggingFace
2

iPrefBench: A User-Centric Benchmark for Personalized Mobile Use Agents

Yuting Ning, Ruijia Cheng, Sushobhan Ghosh, Peitong Duan, HungShi Lin

Apple

Semantic Concept Steering Breaks the Explanation Drift Loop in Continual Learning
HuggingFace
1

Semantic Concept Steering Breaks the Explanation Drift Loop in Continual Learning

Yehonatan Elisha, Oren Barkan, Noam Koenigstein

Tel Aviv University

Retrieval
ThinkBit: Search Codewords, Recalibrate Scales, Select by Reasoning
HuggingFace
0

ThinkBit: Search Codewords, Recalibrate Scales, Select by Reasoning

Rhea Sanjay Sukthanker, Pashmina Cameron

Microsoft

NLP
Gluing Local Contexts into Global Meaning: A Sheaf-Theoretic Decomposition of Transformer Representations
HuggingFace
0

Gluing Local Contexts into Global Meaning: A Sheaf-Theoretic Decomposition of Transformer Representations

Bryce Grant, Peng Wang

Computer Vision
Learning from Automatic 3D Gaussian Splatting Edits for Online Bayesian Scene Change Detection
HuggingFace
0

Learning from Automatic 3D Gaussian Splatting Edits for Online Bayesian Scene Change Detection

Changrui Zhu, Simon Julier

AI Research Papers — October 27, 2026 Edition | Agentic Brew