What do your logits know?

arXiv:2604.09885 · cs.AI · Submitted 2026-04-10 · Read on arXiv

cs.AI

Submitted: 2026-04-10

Updated: 2026-10-05

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Information Leakage — 1× in this paper · explained in 2 episodes
• Vision-language models — 1× in this paper · explained in 2 episodes

Sources

• Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond
• Discovering Latent Knowledge in Language Models Without Supervision
• Eliciting Secret Knowledge from Language Models
• Language Models Represent Space and Time
• Overthinking the Truth: Understanding how Language Models Process False Demonstrations
• CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning
• Evaluating Object Hallucination in Large Vision-Language Models
• Microsoft COCO: Common Objects in Context
• The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
• LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations
• Opening the Black Box of Deep Neural Networks via Information
• BERT Rediscovers the Classical NLP Pipeline
• Deep Learning and the Information Bottleneck Principle
• Breaking the Softmax Bottleneck: A High-Rank RNN Language Model

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems