{"total":53,"rows":[{"id":"2601.11516","title":"Building Production-Ready Probes For Gemini","created":"2026-01-16","cited_by_count":47,"area":"safety"},{"id":"2601.04603","title":"Constitutional Classifiers++: Efficient Production-Grade Defenses against Universal Jailbreaks","created":"2026-01-08","cited_by_count":41,"area":"safety"},{"id":"2603.05488","title":"Reasoning Theater: Disentangling Model Beliefs from Chain-of-Thought","created":"2026-03-05","cited_by_count":40,"area":"lr"},{"id":"2509.26238","title":"Beyond Linear Probes: Dynamic Safety Monitoring for Language Models","created":"2025-09-30","cited_by_count":16,"area":"safety"},{"id":"2602.23783","title":"Diffusion Probe: Generated Image Result Prediction Using CNN Probes","created":"2026-02-27","cited_by_count":14,"area":"diffusion"},{"id":"2508.15882","title":"Beyond Transcription: Mechanistic Interpretability in ASR","created":"2025-08-21","cited_by_count":13,"area":"mechinterp"},{"id":"2511.22662","title":"Difficulties with Evaluating a Deception Detector for AIs","created":"2025-11-27","cited_by_count":13,"area":null},{"id":"2507.15886","title":"Combining Cost-Constrained Runtime Monitors for AI Safety","created":"2025-07-19","cited_by_count":10,"area":null},{"id":"2511.17408","title":"The Impact of Off-Policy Training Data on Probe Generalisation","created":"2025-11-21","cited_by_count":9,"area":null},{"id":"2512.11949","title":"Neural Chameleons: Language Models Can Learn to Hide Their Thoughts from Unseen Activation Monitors","created":"2025-12-12","cited_by_count":9,"area":"lr"},{"id":"2609.29429","title":"Just Ask Jev: Reinforcement Learning for Calibrated Decisions as a Zero-Shot Detector of AI Alignment Failures","created":"2026-09-24","cited_by_count":8,"area":"safety"},{"id":"2604.01151","title":"Detecting Multi-Agent Collusion Through Multi-Agent Interpretability","created":"2026-04-01","cited_by_count":7,"area":"agents"},{"id":"2604.03058","title":"Verbalizing LLMs' assumptions to explain and control sycophancy","created":"2026-04-03","cited_by_count":5,"area":null},{"id":"2605.18549","title":"Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning Dynamics","created":"2026-05-18","cited_by_count":5,"area":"lr"},{"id":"2508.05625","title":"How Do LLMs Persuade? Linear Probes Can Uncover Persuasion Dynamics in Multi-Turn Conversations","created":"2025-08-07","cited_by_count":4,"area":"mechinterp"},{"id":"2605.29601","title":"Training Deliberative Monitors for Black-Box Scheming Detection","created":"2026-05-28","cited_by_count":4,"area":null},{"id":"2512.02282","title":"DialogGuard: Multi-Agent Psychosocial Safety Evaluation of Sensitive LLM Responses","created":"2025-12-01","cited_by_count":3,"area":"agents"},{"id":"2604.14251","title":"Calibrate-Then-Delegate: Safety Monitoring with Risk and Budget Guarantees via Model Cascades","created":"2026-04-15","cited_by_count":3,"area":"safety"},{"id":"2604.14865","title":"Segment-Level Coherence for Robust Harmful Intent Probing in LLMs","created":"2026-04-16","cited_by_count":3,"area":"safety"},{"id":"2605.06890","title":"Beyond the Black Box: Interpretability of Agentic AI Tool Use","created":"2026-05-07","cited_by_count":3,"area":"agents"},{"id":"2606.16364","title":"Looking Is Not Picking: An Attention-Segment Account of Tool-Selection Failures in LLM Agents","created":"2026-06-15","cited_by_count":3,"area":"agents"},{"id":"2606.22673","title":"AgentLens: Interpretable Safety Steering via Mechanistic Subspaces for Multi-Turn Coding Agent","created":"2026-06-21","cited_by_count":3,"area":"agents"},{"id":"2509.21344","title":"Linear probes rely on textual evidence: Results from leakage mitigation studies in language models","created":"2025-09-16","cited_by_count":2,"area":"mechinterp"},{"id":"2606.12360","title":"Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal","created":"2026-06-10","cited_by_count":2,"area":null},{"id":"2607.13075","title":"The Entanglement Wall: Activation-Space Probes as Risk Detectors, Not Context Adjudicators","created":"2026-07-12","cited_by_count":2,"area":"safety"},{"id":"2609.19101","title":"Monitoring and Discovering Reward Hacking with Internal Representations during LLM Evaluations","created":"2026-09-16","cited_by_count":2,"area":"alignment"},{"id":"2511.00554","title":"Red-teaming Activation Probes using Prompted LLMs","created":"2025-11-01","cited_by_count":1,"area":"safety"},{"id":"2601.14660","title":"NeuroFilter: Activation-Based Guardrails for Privacy-Conscious LLM Agents","created":"2026-01-21","cited_by_count":1,"area":"agents"},{"id":"2604.11662","title":"A Robust Evaluation of Probe Robustness: Lessons for Reliable OOD Uncertainty Quantification","created":"2026-04-13","cited_by_count":1,"area":null},{"id":"2608.07139","title":"Online Conformal Prediction Beyond Feedback","created":"2026-08-07","cited_by_count":1,"area":null},{"id":"2609.36490","title":"LLMs Learn to Evade Latent Monitors from Prior Feedback Alone","created":"2026-09-29","cited_by_count":1,"area":"agents"},{"id":"2605.11093","title":"Enabling Performant and Flexible Model-Internal Observability for LLM Inference","created":"2026-05-11","cited_by_count":0,"area":null},{"id":"2605.25893","title":"$D^2$-Monitor: Dynamic Safety Monitoring for Diffusion LLMs via Hesitation Signals","created":"2026-05-25","cited_by_count":0,"area":null},{"id":"2606.07631","title":"Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning","created":"2026-05-31","cited_by_count":0,"area":null},{"id":"2607.14070","title":"Screening of Biosecurity Features in Metagenomic Data with Evo 2 Probes","created":"2026-07-15","cited_by_count":0,"area":null},{"id":"2608.04928","title":"Does Out-of-Sight Equal Out-of-Mind in CoT Monitorability?","created":"2026-08-05","cited_by_count":0,"area":"lr"},{"id":"2608.09643","title":"Activation Probes Surface Code-Security Signals that the Model's Output Misses","created":"2026-08-10","cited_by_count":0,"area":"agents"},{"id":"2608.10621","title":"ProbGuard: Calibrated Safety Risk Estimation from LLM Output Distributions","created":"2026-08-11","cited_by_count":0,"area":"safety"},{"id":"2608.20347","title":"Who Do Language Models Think Is Competent? A Mechanistic Analysis of Occupational Bias","created":"2026-06-15","cited_by_count":0,"area":"lr"},{"id":"2609.00519","title":"The Safeguard Worked. Is the LLM System Safer?","created":"2026-09-01","cited_by_count":0,"area":null},{"id":"2609.00904","title":"In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access?","created":"2026-09-01","cited_by_count":0,"area":null},{"id":"2609.05797","title":"Safety Monitors Mostly Catch What the Model Already Refuses","created":"2026-09-05","cited_by_count":0,"area":"safety"},{"id":"2609.17327","title":"Vroom-Vroom at SHROOM-Visions: A Multi-Judge Committee for Detecting Hallucinated Spans in Vision-Language Outputs","created":"2026-09-15","cited_by_count":0,"area":"mllm"},{"id":"2609.31796","title":"Same Probe, Different Numbers: Are Activation Probes Robust to Inference-Time Numerical Non-Determinism?","created":"2026-09-25","cited_by_count":0,"area":null},{"id":"2609.34771","title":"When Do Model Internals Help? Exploring the Role of Representation Engineering in LLM Safety","created":"2026-09-28","cited_by_count":0,"area":"alignment"},{"id":"2610.00348","title":"Removing the NEEDLE in the Haystack: Backdoor Removal in LLMs via Weight Orthogonalisation","created":"2026-09-29","cited_by_count":0,"area":"safety"},{"id":"2610.02413","title":"Prompted to Discriminate: Generalizing Malicious-Input Probes in the Wild","created":"2026-10-01","cited_by_count":0,"area":"agents"},{"id":"2610.04575","title":"From Probe Scores to Alarm Policies: Operational Validity of Activation Monitors for Language-Model Agents","created":"2026-10-03","cited_by_count":0,"area":"agents"},{"id":"2610.04737","title":"PyINE: A Framework for Scalable Elicitation and Oversight via Code Execution","created":"2026-10-03","cited_by_count":0,"area":"lr"},{"id":"2610.10594","title":"Coverage, Not Difficulty, Sets How Much Synthetic Data an Activation Probe Needs","created":"2026-10-06","cited_by_count":0,"area":"synth"},{"id":"2610.12445","title":"Caught in the Act: Probes Effectively Detect Sabotage and Catch Unverbalized Deception","created":"2026-10-08","cited_by_count":0,"area":"agents"},{"id":"2610.15314","title":"Accuracy and Uncertainty in Machine Deception Detection (Research White Paper)","created":"2026-10-11","cited_by_count":0,"area":null},{"id":"2610.18408","title":"Prompt-Injection Probes Learn the Conversation, Not Just the Injection","created":"2026-10-13","cited_by_count":0,"area":"safety"}],"next_cursor":null}