·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
The Curse of Multilinguality in Lexical Normalization8h◆OpenAgentFlow: Enabling System-Wide Safety Boundaries for Heterogeneous AI Agent Fleets8h◆Residual Sparsification via Output Importance for Compressing Mixture-of-Experts LLMs8h◆Control-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs8h◆A Closed-Loop Evaluation of Capability Loss and Recovery in Compressed Driving Policies8h◆SOVER: Formal Certification of Optimization Reformulations via LLM-Assisted SMT Verification8h◆Escaping Redundant Reasoning: Structure-Aware Search for Inference-Time LLMs8h◆ContextPipe: Database-Inspired Context Assembly for Long-Horizon Agents8h◆TempCloze: Can Video-LLMs Identify the Missing Middle?8h◆Reinforcement Learning Enhanced LLM Agents for Complex Vehicle Routing Problems8h◆Agentic Empirical Asset Pricing: Methodological Foundations8h◆Multi-Step Knowledge Interaction Analysis via Rank-2 Subspace Disentanglement8h◆Can LLMs Discover Scientific Laws in Real and Parallel Worlds?8h◆Consistency Without Alignment: Item-Sensitive Language Models Indistinguishable From Random8h◆Automated Tree Knowledge Graph Construction using Ontology Expansion and Retrieval from Vietnamese History Textbooks8h◆Workload Identification with Physical Side Channels for AI Governance8h◆CoLT-Drive: Counterfactual Long-Tail Benchmarking and Knowledge-Preserving Adaptation for Driving Affordance Prediction8h◆ALEE: Any-Language Evaluation of Embeddings via English-Centric Minimal Pairs8h◆Different representation learning objectives recover distinct latent structures from the same psychometric data8h◆Hidden State Poisoning Attacks against Mamba-based Language Models8h◆The Curse of Multilinguality in Lexical Normalization8h◆OpenAgentFlow: Enabling System-Wide Safety Boundaries for Heterogeneous AI Agent Fleets8h◆Residual Sparsification via Output Importance for Compressing Mixture-of-Experts LLMs8h◆Control-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs8h◆A Closed-Loop Evaluation of Capability Loss and Recovery in Compressed Driving Policies8h◆SOVER: Formal Certification of Optimization Reformulations via LLM-Assisted SMT Verification8h◆Escaping Redundant Reasoning: Structure-Aware Search for Inference-Time LLMs8h◆ContextPipe: Database-Inspired Context Assembly for Long-Horizon Agents8h◆TempCloze: Can Video-LLMs Identify the Missing Middle?8h◆Reinforcement Learning Enhanced LLM Agents for Complex Vehicle Routing Problems8h◆Agentic Empirical Asset Pricing: Methodological Foundations8h◆Multi-Step Knowledge Interaction Analysis via Rank-2 Subspace Disentanglement8h◆Can LLMs Discover Scientific Laws in Real and Parallel Worlds?8h◆Consistency Without Alignment: Item-Sensitive Language Models Indistinguishable From Random8h◆Automated Tree Knowledge Graph Construction using Ontology Expansion and Retrieval from Vietnamese History Textbooks8h◆Workload Identification with Physical Side Channels for AI Governance8h◆CoLT-Drive: Counterfactual Long-Tail Benchmarking and Knowledge-Preserving Adaptation for Driving Affordance Prediction8h◆ALEE: Any-Language Evaluation of Embeddings via English-Centric Minimal Pairs8h◆Different representation learning objectives recover distinct latent structures from the same psychometric data8h◆Hidden State Poisoning Attacks against Mamba-based Language Models8h◆
News/A Causal Model for Locating and Unlocking Sandbagging in Model Organisms
arxiv
PublishedSeptember 1, 2026 at 4:00 AM

A Causal Model for Locating and Unlocking Sandbagging in Model Organisms

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

arXiv:2608.29461v1 Announce Type: new Abstract: Sandbagging models strategically underperform on evaluations while retaining the capabilities being measured. The evaluations that guide frontier-model deployment and governance then understate what these models can do. To understand the mechanism, we

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
arxivThe Curse of Multilinguality in Lexical Normalization8harxivOpenAgentFlow: Enabling System-Wide Safety Boundaries for Heterogeneous AI Agent Fleets8harxivResidual Sparsification via Output Importance for Compressing Mixture-of-Experts LLMs8harxivControl-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs8h
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