{"total":63,"rows":[{"id":"2409.01586","title":"Booster: Tackling Harmful Fine-tuning for Large Language Models via Attenuating Harmful Perturbation","created":"2024-09-03","cited_by_count":92,"area":null},{"id":"2410.07471","title":"SEAL: Safety-enhanced Aligned LLM Fine-tuning via Bilevel Data Selection","created":"2024-10-09","cited_by_count":64,"area":null},{"id":"2403.15447","title":"Decoding Compressed Trust: Scrutinizing the Trustworthiness of Efficient LLMs Under Compression","created":"2024-03-18","cited_by_count":57,"area":null},{"id":"2406.06144","title":"Language Models Resist Alignment: Evidence From Data Compression","created":"2024-06-10","cited_by_count":53,"area":null},{"id":"2412.07097","title":"On Evaluating the Durability of Safeguards for Open-Weight LLMs","created":"2024-12-10","cited_by_count":52,"area":null},{"id":"2406.18664","title":"Evaluating Copyright Takedown Methods for Language Models","created":"2024-06-26","cited_by_count":43,"area":null},{"id":"2502.05209","title":"Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilities","created":"2025-02-03","cited_by_count":39,"area":null},{"id":"2402.16382","title":"Immunization against harmful fine-tuning attacks","created":"2024-02-26","cited_by_count":35,"area":null},{"id":"2501.18100","title":"Panacea: Mitigating Harmful Fine-tuning for Large Language Models via Post-fine-tuning Perturbation","created":"2025-01-30","cited_by_count":20,"area":null},{"id":"2506.08473","title":"AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin","created":"2025-06-10","cited_by_count":19,"area":null},{"id":"2405.19524","title":"AI Risk Management Should Incorporate Both Safety and Security","created":"2024-05-29","cited_by_count":17,"area":null},{"id":"2406.02378","title":"On the Intrinsic Self-Correction Capability of LLMs: Uncertainty and Latent Concept","created":"2024-06-04","cited_by_count":16,"area":null},{"id":"2502.02260","title":"Position: Adversarial ML for LLMs Is Not Making Any Progress","created":"2025-02-04","cited_by_count":15,"area":null},{"id":"2508.17158","title":"Towards Safeguarding LLM Fine-tuning APIs against Cipher Attacks","created":"2025-08-23","cited_by_count":11,"area":null},{"id":"2403.03942","title":"The Heuristic Core: Understanding Subnetwork Generalization in Pretrained Language Models","created":"2024-03-06","cited_by_count":10,"area":null},{"id":"2508.09853","title":"STREAM (ChemBio): A Standard for Transparently Reporting Evaluations in AI Model Reports","created":"2025-08-13","cited_by_count":10,"area":null},{"id":"2410.04524","title":"Toward Secure Tuning: Mitigating Security Risks from Instruction Fine-Tuning","created":"2024-10-06","cited_by_count":8,"area":null},{"id":"2502.00840","title":"Activation Approximations Can Incur Safety Vulnerabilities Even in Aligned LLMs: Comprehensive Analysis and Defense","created":"2025-02-02","cited_by_count":7,"area":null},{"id":"2502.16770","title":"LED-Merging: Mitigating Safety-Utility Conflicts in Model Merging with Location-Election-Disjoint","created":"2025-02-24","cited_by_count":7,"area":null},{"id":"2509.06795","title":"Anchoring Refusal Direction: Mitigating Safety Risks in Tuning via Projection Constraint","created":"2025-09-08","cited_by_count":7,"area":null},{"id":"2510.07985","title":"Fewer Weights, More Problems: A Practical Attack on LLM Pruning","created":"2025-10-09","cited_by_count":7,"area":null},{"id":"2406.10576","title":"Bypass Back-propagation: Optimization-based Structural Pruning for Large Language Models via Policy Gradient","created":"2024-06-15","cited_by_count":6,"area":null},{"id":"2409.03274","title":"Recent Advances in Attack and Defense Approaches of Large Language Models","created":"2024-09-05","cited_by_count":6,"area":null},{"id":"2502.12055","title":"Designing Role Vectors to Improve LLM Inference Behaviour","created":"2025-02-17","cited_by_count":6,"area":null},{"id":"2410.16672","title":"The Tug of War Within: Mitigating the Fairness-Privacy Conflicts in Large Language Models","created":"2024-10-22","cited_by_count":5,"area":null},{"id":"2502.06809","title":"Neurons Speak in Ranges: Breaking Free from Discrete Neuronal Attribution","created":"2025-02-04","cited_by_count":5,"area":null},{"id":"2505.23724","title":"SC-LoRA: Balancing Efficient Fine-tuning and Knowledge Preservation via Subspace-Constrained LoRA","created":"2025-05-29","cited_by_count":5,"area":null},{"id":"2511.07482","title":"Alignment-Constrained Dynamic Pruning for LLMs: Identifying and Preserving Alignment-Critical Circuits","created":"2025-11-09","cited_by_count":4,"area":null},{"id":"2601.22313","title":"Hair-Trigger Alignment: Black-Box Evaluation Cannot Guarantee Post-Update Alignment","created":"2026-01-29","cited_by_count":4,"area":null},{"id":"2503.03874","title":"LEWIS (LayEr WIse Sparsity) -- A Training Free Guided Model Merging Approach","created":"2025-03-05","cited_by_count":3,"area":null},{"id":"2506.05610","title":"Mitigating Confounding in Speech-Based Dementia Detection through Weight Masking","created":"2025-06-05","cited_by_count":3,"area":null},{"id":"2510.08734","title":"Transmuting prompts into weights","created":"2025-10-09","cited_by_count":3,"area":null},{"id":"2512.01899","title":"Provably Safe Model Updates","created":"2025-12-01","cited_by_count":3,"area":null},{"id":"2502.07771","title":"Breaking Down Bias: On The Limits of Generalizable Pruning Strategies","created":"2025-02-11","cited_by_count":2,"area":null},{"id":"2512.22671","title":"Fragile Knowledge, Robust Instruction-Following: The Width Pruning Dichotomy in Llama-3.2","created":"2025-12-27","cited_by_count":2,"area":null},{"id":"2605.08137","title":"Weight Pruning Amplifies Bias: A Multi-Method Study of Compressed LLMs for Edge AI","created":"2026-05-02","cited_by_count":2,"area":null},{"id":"2605.14605","title":"A Few Steps Further: Why Defenses Against Malicious Finetuning Erode Under Continued Training","created":"2026-05-14","cited_by_count":2,"area":null},{"id":"2606.22676","title":"Skin-Deep: A Geometric Diagnostic for Alignment Fragility in Large Language Model Representations","created":"2026-06-21","cited_by_count":2,"area":null},{"id":"2607.02047","title":"OpenSafeIntent: Evaluating Intent-Calibrated Safe Completion Across Dual-Use Prompt Sets","created":"2026-07-02","cited_by_count":2,"area":null},{"id":"2608.28327","title":"Layered LLM Defenses as an Ensemble: Access Tiers, Inference Cost, and the Measured Failure Correlation Between Defense Layers","created":"2026-08-28","cited_by_count":2,"area":null},{"id":"2406.10415","title":"PRISM: A Design Framework for Open-Source Foundation Model Safety","created":"2024-06-14","cited_by_count":1,"area":null},{"id":"2412.20476","title":"Cut the Deadwood Out: Backdoor Purification via Guided Module Substitution","created":"2024-12-29","cited_by_count":1,"area":null},{"id":"2502.06560","title":"Position: It's Time to Act on the Risk of Efficient Personalized Text Generation","created":"2025-02-10","cited_by_count":1,"area":null},{"id":"2604.15294","title":"How Do LLMs and VLMs Understand Viewpoint Rotation Without Vision? An Interpretability Study","created":"2026-04-16","cited_by_count":1,"area":null},{"id":"2606.00160","title":"DataShield: Safety-degrading Data Filtering for LLM Benign Instruction Fine-Tuning","created":"2026-05-29","cited_by_count":1,"area":null},{"id":"2606.10154","title":"Quality Is Not a Safety Proxy Under Quantization","created":"2026-06-08","cited_by_count":1,"area":null},{"id":"2504.04238","title":"Sensitivity Meets Sparsity: The Impact of Extremely Sparse Parameter Patterns on Theory-of-Mind of Large Language Models","created":"2025-04-05","cited_by_count":0,"area":null},{"id":"2601.10524","title":"Diagnosing Generalization Failures in Fine-Tuned LLMs: A Cross-Architectural Study on Phishing Detection","created":"2026-01-15","cited_by_count":0,"area":null},{"id":"2602.17546","title":"Learning to Stay Safe: Adaptive Regularization Against Safety Degradation during Fine-Tuning","created":"2026-02-19","cited_by_count":0,"area":null},{"id":"2603.08275","title":"AdaCultureSafe: Adaptive Cultural Safety Grounded by Cultural Knowledge in Large Language Models","created":"2026-03-09","cited_by_count":0,"area":null},{"id":"2604.00324","title":"The Persistent Vulnerability of Aligned AI Systems","created":"2026-03-31","cited_by_count":0,"area":null},{"id":"2605.10575","title":"Acceptance Cards:A Four-Diagnostic Standard for Safe Fine-Tuning Defense Claims","created":"2026-05-11","cited_by_count":0,"area":null},{"id":"2606.03648","title":"Safety Measurements for Fine-tuned LLMs Should be Grounded in Capability","created":"2026-06-02","cited_by_count":0,"area":null},{"id":"2606.05613","title":"Multilingual Fine-Tuning via Localized Gradient Conflict Resolution","created":"2026-06-04","cited_by_count":0,"area":null},{"id":"2606.13705","title":"When Can One Neuron Fix Repetition Loops in LLMs?","created":"2026-06-09","cited_by_count":0,"area":null},{"id":"2606.30219","title":"EvalSafetyGap: A Hybrid Survey and Conceptual Framework for LLM Evaluation-Safety Failures","created":"2026-06-29","cited_by_count":0,"area":null},{"id":"2607.05356","title":"ReCal3R: Reliability-Calibrated Learning Rates for Streaming 3D Reconstruction","created":"2026-07-06","cited_by_count":0,"area":null},{"id":"2609.10142","title":"Active Adaptation, Not Static Defense: Temporal Dynamics of Preventative Steering in Adversarial Fine-Tuning","created":"2026-09-09","cited_by_count":0,"area":null},{"id":"2609.22153","title":"SafeTune: A Unified Faithful Library for Auditing and Repairing Safety Drift in Fine-Tuned LLMs","created":"2026-08-26","cited_by_count":0,"area":null},{"id":"2609.35117","title":"Tool Mediation Alters Refusal Mechanisms in Large Language Models","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35261","title":"Imprint Reader: From Weight-Update Readout to Behavioral Intervention","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2610.13928","title":"AI Models Can Provably Hide Arbitrary Capabilities","created":"2026-10-09","cited_by_count":0,"area":null},{"id":"2610.18098","title":"From Spectral Saliency to Finite Deletion: Evaluating Selective Language Model Edits","created":"2026-10-13","cited_by_count":0,"area":null}],"next_cursor":null}