·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Enhancing Autonomous Online Intrusion Detection for IoT with Balanced Learning, Reliable Pseudo-Labels, and Lightweight Architectures7h◆PitchBench: Measuring Pitch Hearing in Audio-Language Models7h◆JobBench: Aligning Agent Work With Human Will7h◆AnchorDiff: Training-Free Concept Grounding for MM-DiTs via Anchor-Based Graph Propagation7h◆GeoFaith: A Spatio-Temporal Dual View of Faithful Chain-of-Thought7h◆Eroding Trust in Real Speech: A Large-Scale Study of Human Audio Deepfake Perception7h◆Yes, Q-learning Helps Offline In-Context RL7h◆Telenor Nordics Customer Service self-help corpus7h◆Stylistic Evolution and LLM Neutrality in Singlish Language7h◆SPHERE-JEPA: Spherical Prediction with Homogeneous Embeddings7h◆LearnedCache: An eBPF-Integrated Perceptron-Based Eviction Policy for the Linux Page Cache7h◆Efficient Learning of Mesh-Based Physical Simulation with BSMS-GNN7h◆Sample Complexity of Policy Gradient for Log-Growth Control7h◆Proper Calibeating7h◆Interpretability and Generalization Bounds for Learning Spatial Physics7h◆Constrained Meta Reinforcement Learning with Provable Test-Time Safety7h◆Step-Size Stability in Stochastic Optimization: A Theoretical Perspective7h◆Two-Parameter Flows for Learning Population Dynamics of Physical Systems7h◆Stateful Inference for Low-Latency Multi-Agent Tool Calling7h◆Separate Aggregation of Split Network for Personalized Federated Learning7h◆Enhancing Autonomous Online Intrusion Detection for IoT with Balanced Learning, Reliable Pseudo-Labels, and Lightweight Architectures7h◆PitchBench: Measuring Pitch Hearing in Audio-Language Models7h◆JobBench: Aligning Agent Work With Human Will7h◆AnchorDiff: Training-Free Concept Grounding for MM-DiTs via Anchor-Based Graph Propagation7h◆GeoFaith: A Spatio-Temporal Dual View of Faithful Chain-of-Thought7h◆Eroding Trust in Real Speech: A Large-Scale Study of Human Audio Deepfake Perception7h◆Yes, Q-learning Helps Offline In-Context RL7h◆Telenor Nordics Customer Service self-help corpus7h◆Stylistic Evolution and LLM Neutrality in Singlish Language7h◆SPHERE-JEPA: Spherical Prediction with Homogeneous Embeddings7h◆LearnedCache: An eBPF-Integrated Perceptron-Based Eviction Policy for the Linux Page Cache7h◆Efficient Learning of Mesh-Based Physical Simulation with BSMS-GNN7h◆Sample Complexity of Policy Gradient for Log-Growth Control7h◆Proper Calibeating7h◆Interpretability and Generalization Bounds for Learning Spatial Physics7h◆Constrained Meta Reinforcement Learning with Provable Test-Time Safety7h◆Step-Size Stability in Stochastic Optimization: A Theoretical Perspective7h◆Two-Parameter Flows for Learning Population Dynamics of Physical Systems7h◆Stateful Inference for Low-Latency Multi-Agent Tool Calling7h◆Separate Aggregation of Split Network for Personalized Federated Learning7h◆
News/Do LLMs Experience an Internal Polylogue? Investigating Reasoning through the Lens of Personas
arxiv
PublishedMay 13, 2026 at 4:00 AM
—neutral

Do LLMs Experience an Internal Polylogue? Investigating Reasoning through the Lens of Personas

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

arXiv:2605.09159v1 Announce Type: new Abstract: Recent work shows that large language models (LLMs) encode behavioural traits ("personas") as linear directions in activation space, often called "persona vectors". Prior work has used such directions as static handles for behavioural steering. Buildin

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
arxivEnhancing Autonomous Online Intrusion Detection for IoT with Balanced Learning, Reliable Pseudo-Labels, and Lightweight Architectures7harxivPitchBench: Measuring Pitch Hearing in Audio-Language Models7harxivJobBench: Aligning Agent Work With Human Will7harxivAnchorDiff: Training-Free Concept Grounding for MM-DiTs via Anchor-Based Graph Propagation7h
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