{"total":33,"rows":[{"id":"2510.25771","title":"Gaperon: A Peppered English-French Generative Language Model Suite","created":"2025-10-29","cited_by_count":9,"area":null},{"id":"2605.23645","title":"Learning Through Noise: Why Subliminal Learning Works and When It Fails","created":"2026-05-22","cited_by_count":7,"area":null},{"id":"2601.15220","title":"Privacy Collapse: Benign Fine-Tuning Can Break Contextual Privacy in Language Models","created":"2026-01-21","cited_by_count":6,"area":null},{"id":"2602.09987","title":"Infusion: Shaping Model Behavior by Editing Training Data via Influence Functions","created":"2026-02-10","cited_by_count":5,"area":null},{"id":"2604.13076","title":"Alignment midtraining for animals","created":"2026-03-21","cited_by_count":4,"area":null},{"id":"2602.16729","title":"Intent Laundering: AI Safety Datasets Are Not What They Seem","created":"2026-02-17","cited_by_count":3,"area":null},{"id":"2604.16423","title":"Shifting the Gradient: Understanding How Defensive Training Methods Protect Language Model Integrity","created":"2026-04-03","cited_by_count":3,"area":null},{"id":"2604.23593","title":"When AI reviews science: Can we trust the referee?","created":"2026-04-26","cited_by_count":2,"area":null},{"id":"2510.26829","title":"Layer of Truth: Probing Belief Shifts under Continual Pre-Training Poisoning","created":"2025-10-29","cited_by_count":1,"area":null},{"id":"2602.03899","title":"Byzantine Machine Learning: MultiKrum and an optimal notion of robustness","created":"2026-02-03","cited_by_count":1,"area":null},{"id":"2602.17729","title":"Stop Saying \"AI\"","created":"2026-02-18","cited_by_count":1,"area":null},{"id":"2607.15267","title":"Pretraining Data Can Be Poisoned through Computational Propaganda","created":"2026-07-16","cited_by_count":1,"area":null},{"id":"2607.23394","title":"Inference-Time Consensus for Mitigating Hidden Behaviors from LLM Fine-Tuning","created":"2026-07-25","cited_by_count":1,"area":null},{"id":"2609.17380","title":"OPEN-1B: A Fully Auditable Training Run","created":"2026-09-15","cited_by_count":1,"area":null},{"id":"2511.02600","title":"On The Dangers of Poisoned LLMs In Security Automation","created":"2025-11-04","cited_by_count":0,"area":null},{"id":"2511.13753","title":"Robustness of LLM-enabled vehicle trajectory prediction under data security threats","created":"2025-11-14","cited_by_count":0,"area":null},{"id":"2512.02043","title":"Mirror, Mirror on the Wall -- Which is the Best Model of Them All?","created":"2025-11-25","cited_by_count":0,"area":null},{"id":"2603.07401","title":"VIVECaption: A Split Approach to Caption Quality Improvement","created":"2026-03-08","cited_by_count":0,"area":null},{"id":"2604.25639","title":"Large language models eroding science understanding: an empirical study of malignment","created":"2026-04-28","cited_by_count":0,"area":null},{"id":"2605.16325","title":"Phase Transitions in Driven Informational Systems: A Two-Field Perspective on Learning Theory and Non-Equilibrium Chemistry","created":"2026-05-05","cited_by_count":0,"area":null},{"id":"2605.22481","title":"When Stronger Triggers Backfire: A High-Dimensional Theory of Backdoor Attacks","created":"2026-05-21","cited_by_count":0,"area":null},{"id":"2606.05919","title":"Finding Most Influential Sets","created":"2026-06-04","cited_by_count":0,"area":null},{"id":"2606.07245","title":"AI Sovereignty: A Qualitative Model of Strategic Competition as AI Becomes an Instrument of National Power","created":"2026-06-05","cited_by_count":0,"area":null},{"id":"2607.06643","title":"The Power of Backdoor Absorption in Community Training","created":"2026-07-07","cited_by_count":0,"area":null},{"id":"2608.27954","title":"Not to Break, but to Attest: Adversarial Probes for Privacy-Preserving LLM Verification","created":"2026-08-28","cited_by_count":0,"area":null},{"id":"2609.10883","title":"Story Imprinting: AI Assistants Absorb Traits from Human Characters They Resemble","created":"2026-09-09","cited_by_count":0,"area":null},{"id":"2609.18560","title":"The evolution of sex for artificial intelligence: a population-genetic framework for multigenerational model populations","created":"2026-09-16","cited_by_count":0,"area":null},{"id":"2609.33915","title":"Training Witnesses: Trusting the Training without Trusting the Trainer","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.33985","title":"The Privacy Fallacy of Crowdsourced Fine-Tuning: Extracting Proprietary Data via Topic-Based Poisoning","created":"2026-09-27","cited_by_count":0,"area":null},{"id":"2609.36117","title":"Why Backdooring Neural Networks is so Easy?","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2610.02886","title":"Misinformation Without Triggers: From Factual Answers to Downstream Decisions","created":"2026-10-02","cited_by_count":0,"area":null},{"id":"2610.06994","title":"TARE: Weigh a Never-Poisoned Twin Before Reading Backdoor-Defense Costs","created":"2026-10-04","cited_by_count":0,"area":null},{"id":"2610.14487","title":"AlignPoison: Conditional Targeted Poisoning of Generative Recommenders via Decoding-Aligned Prefixes","created":"2026-10-10","cited_by_count":0,"area":null}],"next_cursor":null}