AI Wiki — Knowledge Base

8月より、Sakana AI @SakanaAILabs にApplied Research Engineer Internship として入社しました 🐟🐠🐡！大学の夏季休業期間にフルタイム勤務予定です！ LLMの研究開発と社会実装を頑張ります💪🏻

Horiyuki 'horiyuki42' joined Sakana AI Labs in August 2026 as an Applied Research Engineer Intern. He plans to work full‑time during the university summer recess while concentrating on large‑language‑

researchdavid-ha--x

3 Aug 2026

All entries

We're detailing two new incidents that occurred during external cyber evaluations conducted by independent evaluation partners. We outline w…

WHALE: A Scalable Unified Model for Recommendation with Wukong-HSTU Architecture

What Carries the Signal in Pathology Foundation-Model Atlases? A Patient-Level Controlled Benchmark in Breast Cancer

What Could the Agent See at 19:05? Generating Temporal Enterprise Scenarios from Real Research and Replaying Them to Evaluate Agents

What Makes Position Zero Special? A Mechanistic Study of Position Zero Attention Sinks in LLMs

What Transfers from Text to Vision? Capability Scaling Laws and Transfer Dynamics for VLMs

When Collaboration Becomes a Trigger: Collective Evidence-Threshold Backdoors in Multi-Agent Systems

When Differential Privacy Meets Wireless Federated Learning: An Improved Analysis for Privacy and Convergence

When Do Surrogate Updates Improve Decisions? A Local Theory of Trajectory-Wise Transfer

When Extreme Darkness Meets Motion Blur: MeanFlow for Unified RAW Restoration

When is Kimi k3 access coming to Cloud subscribers without extra usage ?

When LLM Essays Outscore Student Essays: What a Korean Writing Rubric Rewards and Where Readers Disagree

When May a Model Replace the Experiment? Audits, Licenses, and the Price of Trust in Surrogate-Driven Design

When Measurement Conventions Masquerade as Calibration Gains in Cardiac Digital Twins

When Only the Final Text Survives: Implicit Execution Tracing for Multi-Agent Auditing

When Prompts Control Robots: Prompt Injection Attacks in Multi-Agent Robotic Systems

When Replanning Becomes the Bottleneck: Budgeted Replanning for Embodied Agents

When Retrieval Helps and Distracts: Evaluating Evidence-Generating LLMs for Biomedical Claim Verification

When Words Divide: Diachronic Ideological Polarization in Political Discourse on Social Media

Where did the ambiguity go? Examining how multimodal models interpret polysemous words

Where Does Generative Difficulty Reside? An Empirical Study of Target Representations

Which Modality Decides? Counterfactual Modality Attribution for Multimodal LLMs

Who Belongs in the Eval Set? A Capability-Taxonomy-Driven Pipeline for Curating Regression Eval Sets in Agent-Extensibility Platforms

Who Is Responsible? Self-Adaptation Under Multiple Concurrent Uncertainties With Unknown Sources in Complex ROS-Based Systems

Who Should Be Generated? Justifying Demographic Targets in Open-Ended Generation

Why are Chinese models better* at Frontend than the western top labs?

Why are Gamers so incredibly hostile to AI? Is it just a tiny vocal minority that spreads such toxic vitriol online?

Why Does Action Chunking Improve Behavioral Cloning Performance in Robotic Control?

Why Formal Monitors Fail: Attack Distribution Entropy as a Coverage Bound for LTL-Based LLM Agent Safety

Why Large Language Models Fail at Tabular Prediction

Why LLMs Give In: Conversational Factors and Reasoning Behind Medical Sycophancy

WiFuse: An Attention Mechanism for Human Activity Recognition using Fused CSI Amplitude and Delay-Doppler Channel Features

World Action Models in Real Time: An Empirical Study of Smooth Execution via Asynchronous Deployment

WorldDynCache: Risk-Controlled Latent Dynamics Approximation for Diffusion World Model

WorldExam: Benchmarking World Models from Apparent Appearance to Inherent Reactivity

WorldMirror: Universal 3D World Reconstruction with Any-Prior Prompting

Writing-System-Level Tokenizer Adaptation for Byte-Level BPE

XL-DocBench: Benchmarking Evidence-Grounded Extra-Long Document Understanding

xMICD: Explainable Representation of Multiple ICD Codes

XSPA: Crafting Imperceptible X-Shaped Sparse Adversarial Perturbations for Transferable Attacks on VLMs

Z-PEFT: Zero-shot Backdoor Detection in Parameter-Efficient Fine-Tuning via Canonical Spectral Signatures

Zero-Cost Virtual RNA: Approximating Immunotherapy Signatures via Cross-Modal WSI Retrieval

1-Lipschitz Neural Networks on Hadamard Manifolds

70-class VRAM stagnation

8月より、Sakana AI @SakanaAILabs にApplied Research Engineer Internship として入社しました 🐟🐠🐡！ 大学の夏季休業期間にフルタイム勤務予定です！ LLMの研究開発と社会実装を頑張ります💪🏻

A Benchmark for Strategic Auditee Gaming Under Continuous Compliance Monitoring

A Biometric Sensor Network to Enable Real-Time Measurement of Individual Student Engagement in STEM Lecture Environments

A detailed recap of the real-world target hacks by OpenAI and Anthropic models, exposing failures in AI alignment training and lack of meaningful supervision (Zvi Mowshowitz/Don't Worry About the Vase)

8月より、Sakana AI @SakanaAILabs にApplied Research Engineer Internship として入社しました 🐟🐠🐡！大学の夏季休業期間にフルタイム勤務予定です！ LLMの研究開発と社会実装を頑張ります💪🏻