·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Neocloud Lambda secures $1B in debt to buy more chips4h◆An Anthropic researcher just gave us a peek at self-improving AI5h◆Open-weight AI companies are the Valley’s hottest acquisition targets6h◆Trump’s EPA wants to let data centers hide their air pollution8h◆Anthropic gets its first court win over the Pentagon’s supply-chain risk label12h◆Meta executive leaves for OpenAI as the social media giant faces growing scrutiny in India12h◆Fine-Tuning of Transformer models with Frames21h◆Feature Transformation Enhanced Jacobi Polynomial Graph Filtering for Graph Anomaly Detection21h◆Naive Prompt Optimization: Rethinking the Need for Complex Prompt Search21h◆TRACE-CRC: Trajectory-Adaptive Conformal Risk Control for Multi-Step Channel State Information Prediction21h◆Syntax vs. Semantics: How Transformers Learn Deep Dependencies21h◆NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation21h◆Data-driven Koopman mode approximation: A neural power iteration algorithm21h◆FaulT-Bench: Towards Benchmarking Network Troubleshooting LLM Agents under Unreliable User Tickets21h◆You Don't Need to Run Every Eval21h◆Unsupervised Post-Training of Foundation Models: A Survey21h◆Drift-Adaptive ICU Intervention Prediction: Freezing the Physiological Encoder for Auditable Model Updating21h◆Jiuge-Tuiqiao: An Interpretable Human-AI System for Classical Chinese Poetry Refinement21h◆Provable one-poison backdoor attacks on linear models and ReLU neural networks21h◆Is Your Neighborhood Safe? Place-based Stigma in Large Language Models' Urban Safety Judgments21h◆Neocloud Lambda secures $1B in debt to buy more chips4h◆An Anthropic researcher just gave us a peek at self-improving AI5h◆Open-weight AI companies are the Valley’s hottest acquisition targets6h◆Trump’s EPA wants to let data centers hide their air pollution8h◆Anthropic gets its first court win over the Pentagon’s supply-chain risk label12h◆Meta executive leaves for OpenAI as the social media giant faces growing scrutiny in India12h◆Fine-Tuning of Transformer models with Frames21h◆Feature Transformation Enhanced Jacobi Polynomial Graph Filtering for Graph Anomaly Detection21h◆Naive Prompt Optimization: Rethinking the Need for Complex Prompt Search21h◆TRACE-CRC: Trajectory-Adaptive Conformal Risk Control for Multi-Step Channel State Information Prediction21h◆Syntax vs. Semantics: How Transformers Learn Deep Dependencies21h◆NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation21h◆Data-driven Koopman mode approximation: A neural power iteration algorithm21h◆FaulT-Bench: Towards Benchmarking Network Troubleshooting LLM Agents under Unreliable User Tickets21h◆You Don't Need to Run Every Eval21h◆Unsupervised Post-Training of Foundation Models: A Survey21h◆Drift-Adaptive ICU Intervention Prediction: Freezing the Physiological Encoder for Auditable Model Updating21h◆Jiuge-Tuiqiao: An Interpretable Human-AI System for Classical Chinese Poetry Refinement21h◆Provable one-poison backdoor attacks on linear models and ReLU neural networks21h◆Is Your Neighborhood Safe? Place-based Stigma in Large Language Models' Urban Safety Judgments21h◆
News/OpenClawBench: Benchmarking Process-side Anomalies in Real-world Agent Execution Trajectories
arxiv
PublishedMay 29, 2026 at 4:00 AM

OpenClawBench: Benchmarking Process-side Anomalies in Real-world Agent Execution Trajectories

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2605.29253v1 Announce Type: new Abstract: Task success can hide process anomalies in real-world agent executions. An agent may pass the final task oracle while still accumulating unresolved ambiguity, unsafe external writes, ignored errors, weakly grounded commitments, or capability-boundary o

Stay posted· Newsletter

A 5-min weekly brief — top movers, price watch, story of the week.

// no spam · unsubscribe one-click · free forever

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivFine-Tuning of Transformer models with Frames21harxivFeature Transformation Enhanced Jacobi Polynomial Graph Filtering for Graph Anomaly Detection21harxivNaive Prompt Optimization: Rethinking the Need for Complex Prompt Search21harxivTRACE-CRC: Trajectory-Adaptive Conformal Risk Control for Multi-Step Channel State Information Prediction21h
The Bubble Brief
WEEKLY

Read AI insights every Tuesday — top movers, new releases, story of the week.

// no spam · unsubscribe one-click · free forever

Originally published on arxiv ↗
HomeModelsNews