{"total":51,"rows":[{"id":"2303.15180","title":"Detecting Backdoors in Pre-trained Encoders","created":"2023-03-23","cited_by_count":55,"area":"safety"},{"id":"2601.11516","title":"Building Production-Ready Probes For Gemini","created":"2026-01-16","cited_by_count":47,"area":"safety"},{"id":"2401.16765","title":"A Cross-Language Investigation into Jailbreak Attacks in Large Language Models","created":"2024-01-30","cited_by_count":41,"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":"2410.12949","title":"Mechanistic Unlearning: Robust Knowledge Unlearning and Editing via Mechanistic Localization","created":"2024-10-16","cited_by_count":35,"area":"safety"},{"id":"2202.11202","title":"Indiscriminate Poisoning Attacks on Unsupervised Contrastive Learning","created":"2022-02-22","cited_by_count":34,"area":"safety"},{"id":"2411.03343","title":"What Features in Prompts Jailbreak LLMs? Investigating the Mechanisms Behind Attacks","created":"2024-11-02","cited_by_count":29,"area":"safety"},{"id":"2503.10566","title":"ASIDE: Architectural Separation of Instructions and Data in Language Models","created":"2025-03-13","cited_by_count":24,"area":"safety"},{"id":"2411.12701","title":"When Backdoors Speak: Understanding LLM Backdoor Attacks Through Model-Generated Explanations","created":"2024-11-19","cited_by_count":20,"area":"safety"},{"id":"2310.18633","title":"Setting the Trap: Capturing and Defeating Backdoors in Pretrained Language Models through Honeypots","created":"2023-10-28","cited_by_count":19,"area":"safety"},{"id":"2405.04160","title":"A Causal Explainable Guardrails for Large Language Models","created":"2024-05-07","cited_by_count":18,"area":"safety"},{"id":"2603.12277","title":"Prompt Injection as Role Confusion","created":"2026-02-22","cited_by_count":17,"area":"safety"},{"id":"2509.26238","title":"Beyond Linear Probes: Dynamic Safety Monitoring for Language Models","created":"2025-09-30","cited_by_count":16,"area":"safety"},{"id":"2604.18519","title":"LLM Safety From Within: Detecting Harmful Content with Internal Representations","created":"2026-04-20","cited_by_count":11,"area":"safety"},{"id":"2407.17710","title":"Revisiting Machine Unlearning with Dimensional Alignment","created":"2024-07-25","cited_by_count":10,"area":"safety"},{"id":"2411.19876","title":"LUMIA: Linear probing for Unimodal and MultiModal Membership Inference Attacks leveraging internal LLM states","created":"2024-11-29","cited_by_count":10,"area":"safety"},{"id":"2506.05451","title":"Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety","created":"2025-06-05","cited_by_count":10,"area":"safety"},{"id":"2405.16783","title":"TrojFM: Resource-efficient Backdoor Attacks against Very Large Foundation Models","created":"2024-05-27","cited_by_count":5,"area":"safety"},{"id":"2412.13435","title":"Lightweight Safety Classification Using Pruned Language Models","created":"2024-12-18","cited_by_count":5,"area":"safety"},{"id":"2505.20435","title":"The Shape of Adversarial Influence: Characterizing LLM Latent Spaces with Persistent Homology","created":"2025-05-26","cited_by_count":5,"area":"safety"},{"id":"2502.15594","title":"SafeInt: Shielding Large Language Models from Jailbreak Attacks via Safety-Aware Representation Intervention","created":"2025-02-21","cited_by_count":4,"area":"safety"},{"id":"2510.25117","title":"A Survey on Unlearning in Large Language Models","created":"2025-10-29","cited_by_count":4,"area":"safety"},{"id":"2604.15166","title":"Class Unlearning via Depth-Aware Removal of Forget-Specific Directions","created":"2026-04-16","cited_by_count":4,"area":"safety"},{"id":"2404.08631","title":"FCert: Certifiably Robust Few-Shot Classification in the Era of Foundation Models","created":"2024-04-12","cited_by_count":3,"area":"safety"},{"id":"2601.15111","title":"Auditing Language Model Unlearning via Information Decomposition","created":"2026-01-21","cited_by_count":3,"area":"safety"},{"id":"2601.21564","title":"Representation Unlearning: Forgetting through Information Compression","created":"2026-01-29","cited_by_count":2,"area":"safety"},{"id":"2603.20206","title":"Enhancing Safety of Large Language Models via Embedding Space Separation","created":"2026-03-01","cited_by_count":2,"area":"safety"},{"id":"2605.18646","title":"Language-Switching Triggers Take a Latent Detour Through Language Models","created":"2026-05-18","cited_by_count":2,"area":"safety"},{"id":"2608.08027","title":"BASIS: Breach-Aware Selective Prompt Injection Shielding with Prefill Attention Probes","created":"2026-08-08","cited_by_count":2,"area":"safety"},{"id":"2512.14751","title":"One Leak Away: How Pretrained Model Exposure Amplifies Jailbreak Risks in Finetuned LLMs","created":"2025-12-14","cited_by_count":1,"area":"safety"},{"id":"2601.12460","title":"TrojanPraise: Jailbreak LLMs via Benign Fine-Tuning","created":"2026-01-18","cited_by_count":1,"area":"safety"},{"id":"2605.20282","title":"Do Vision Models Truly Forget? New Findings from Representation-Level Certification of Visual Unlearning in Vertical Federated Learning","created":"2026-05-19","cited_by_count":1,"area":"safety"},{"id":"2606.06320","title":"Learning What to Forget: Improving LLM Unlearning via Learned Token-Level Importance","created":"2026-06-04","cited_by_count":1,"area":"safety"},{"id":"2506.11615","title":"Machine Unlearning for Robust DNNs: Attribution-Guided Partitioning and Neuron Pruning in Noisy Environments","created":"2025-06-13","cited_by_count":0,"area":"safety"},{"id":"2508.21448","title":"When Models Refuse: Political Steerability and Feature Richness as Measures of Ideological Depth","created":"2025-08-29","cited_by_count":0,"area":"safety"},{"id":"2605.02958","title":"Tracing the Dynamics of Refusal: Exploiting Latent Refusal Trajectories for Robust Jailbreak Detection","created":"2026-05-02","cited_by_count":0,"area":"safety"},{"id":"2605.10680","title":"Behavioral Guarantees for Proxy-Based Unlearning","created":"2026-05-11","cited_by_count":0,"area":"safety"},{"id":"2605.14347","title":"Exemplar Partitioning for Mechanistic Interpretability","created":"2026-05-14","cited_by_count":0,"area":"safety"},{"id":"2607.05726","title":"Association Restoration Test: Revealing Restorable Shortcuts after Unlearning","created":"2026-07-07","cited_by_count":0,"area":"safety"},{"id":"2607.15218","title":"When Words Are Safe But Actions Kill: Probing Physical Jailbreak Beyond Textual Jailbreak in Hidden-State Risk Space","created":"2026-07-16","cited_by_count":0,"area":"safety"},{"id":"2608.10621","title":"ProbGuard: Calibrated Safety Risk Estimation from LLM Output Distributions","created":"2026-08-11","cited_by_count":0,"area":"safety"},{"id":"2608.12652","title":"Excess Separability: Nuisance-Controlled Residual-Stream Probing for Benchmark Contamination Detection","created":"2026-08-12","cited_by_count":0,"area":"safety"},{"id":"2609.07746","title":"LLM Forensics: Where Do Backdoors Hide? Localizing and Controlling Trigger Mechanisms with Sparse Autoencoders","created":"2026-09-07","cited_by_count":0,"area":"safety"},{"id":"2609.13737","title":"ForeSight: Enhancing Risk Monitoring via Early Safety Signal Distillation","created":"2026-09-12","cited_by_count":0,"area":"safety"},{"id":"2609.21996","title":"A Lie Detector Test for Language Models: Reading Knowledge a Model Won't Reveal","created":"2026-09-18","cited_by_count":0,"area":"safety"},{"id":"2609.31062","title":"CG-Probes: Recovering Guardrail Directions from Patient Query Embeddings","created":"2026-09-25","cited_by_count":0,"area":"safety"},{"id":"2609.36612","title":"Do LLMs Really Forget? Hidden-State Leakage in Model Unlearning and How to Fix it","created":"2026-09-29","cited_by_count":0,"area":"safety"},{"id":"2610.04426","title":"UnAct: Gradient-Free Unlearning via Targeted Activation Intervention","created":"2026-10-03","cited_by_count":0,"area":"safety"},{"id":"2610.05541","title":"Don't Judge an LLM Only by Its Activations: Discovering Suppressed Safety Features via Counterfactual Activation Potential","created":"2026-10-04","cited_by_count":0,"area":"safety"},{"id":"2610.07005","title":"Where Does the Audio Jailbreak Live? A Controlled Frequency-Depth Audit of AdvWave-P on Qwen2-Audio","created":"2026-10-04","cited_by_count":0,"area":"safety"},{"id":"2610.15218","title":"Running Fewer Loops Removes Refusal Behaviour in Recurrent-Depth Language Models","created":"2026-10-10","cited_by_count":0,"area":"safety"}],"next_cursor":null}