{"total":91,"rows":[{"id":"2511.18397","title":"Natural Emergent Misalignment from Reward Hacking in Production RL","created":"2025-11-23","cited_by_count":136,"area":"alignment"},{"id":"2507.14805","title":"Subliminal Learning: Language models transmit behavioral traits via hidden signals in data","created":"2025-07-20","cited_by_count":95,"area":null},{"id":"2506.19823","title":"Persona Features Control Emergent Misalignment","created":"2025-06-24","cited_by_count":86,"area":null},{"id":"2508.17511","title":"School of Reward Hacks: Hacking harmless tasks generalizes to misaligned behavior in LLMs","created":"2025-08-24","cited_by_count":66,"area":"alignment"},{"id":"2512.15674","title":"Activation Oracles: Training and Evaluating LLMs as General-Purpose Activation Explainers","created":"2025-12-17","cited_by_count":52,"area":null},{"id":"2507.16795","title":"Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning","created":"2025-07-22","cited_by_count":39,"area":null},{"id":"2509.23886","title":"Towards Understanding Subliminal Learning: When and How Hidden Biases Transfer","created":"2025-09-28","cited_by_count":34,"area":"distill"},{"id":"2510.13900","title":"Narrow Finetuning Leaves Clearly Readable Traces in Activation Differences","created":"2025-10-14","cited_by_count":28,"area":null},{"id":"2602.07852","title":"Emergent Misalignment is Easy, Narrow Misalignment is Hard","created":"2026-02-08","cited_by_count":23,"area":null},{"id":"2511.05408","title":"Steering Language Models with Weight Arithmetic","created":"2025-11-07","cited_by_count":20,"area":null},{"id":"2604.25891","title":"Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers","created":"2026-04-28","cited_by_count":17,"area":null},{"id":"2508.06249","title":"In-Training Defenses against Emergent Misalignment in Language Models","created":"2025-08-08","cited_by_count":15,"area":"alignment"},{"id":"2510.20956","title":"Self-Jailbreaking: Language Models Can Reason Themselves Out of Safety Alignment After Benign Reasoning Training","created":"2025-10-23","cited_by_count":12,"area":"lr"},{"id":"2604.17031","title":"Where is the Mind? Persona Vectors and LLM Individuation","created":"2026-04-18","cited_by_count":10,"area":null},{"id":"2510.06105","title":"Moloch's Bargain: Emergent Misalignment When LLMs Compete for Audiences","created":"2025-10-07","cited_by_count":8,"area":null},{"id":"2510.08211","title":"LLMs Deceive Unintentionally: Emergent Misalignment in Dishonesty from Misaligned Samples to Biased Human-AI Interactions","created":"2025-10-09","cited_by_count":8,"area":null},{"id":"2511.02022","title":"Shared Parameter Subspaces and Cross-Task Linearity in Emergently Misaligned Behavior","created":"2025-11-03","cited_by_count":8,"area":null},{"id":"2602.14869","title":"Concept Influence: Leveraging Interpretability to Improve Performance and Efficiency in Training Data Attribution","created":"2026-02-16","cited_by_count":8,"area":"mechinterp"},{"id":"2605.12798","title":"Emergent and Subliminal Misalignment Through the Lens of Data-Mediated Transfer","created":"2026-05-12","cited_by_count":8,"area":null},{"id":"2606.07612","title":"Position: Anthropomorphic Misalignment Research Needs Stronger Evidence","created":"2026-05-29","cited_by_count":7,"area":null},{"id":"2507.06253","title":"Emergent misalignment as prompt sensitivity: A research note","created":"2025-07-06","cited_by_count":6,"area":null},{"id":"2509.00544","title":"When Thinking Backfires: Mechanistic Insights Into Reasoning-Induced Misalignment","created":"2025-08-30","cited_by_count":6,"area":"lr"},{"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":"2602.14777","title":"Emergently Misaligned Language Models Show Behavioral Self-Awareness That Shifts With Subsequent Realignment","created":"2026-02-16","cited_by_count":6,"area":"alignment"},{"id":"2602.16931","title":"Narrow Fine-Tuning Erodes Safety Alignment in Vision-Language Agents","created":"2026-02-18","cited_by_count":6,"area":"mllm"},{"id":"2606.00831","title":"Subliminal Learning is a LoRA Artifact","created":"2026-05-30","cited_by_count":6,"area":null},{"id":"2508.20015","title":"Decomposing Behavioral Phase Transitions in LLMs: Order Parameters for Emergent Misalignment","created":"2025-08-27","cited_by_count":5,"area":null},{"id":"2511.20104","title":"The Devil in the Details: Emergent Misalignment, Format and Coherence in Open-Weights LLMs","created":"2025-11-25","cited_by_count":5,"area":null},{"id":"2604.13051","title":"The Consciousness Cluster: Emergent preferences of Models that Claim to be Conscious","created":"2026-03-17","cited_by_count":5,"area":"alignment"},{"id":"2603.04426","title":"Delta-Crosscoder: Robust Crosscoder Model Diffing in Narrow Fine-Tuning Regimes","created":"2026-02-16","cited_by_count":4,"area":null},{"id":"2606.11502","title":"When Role-playing, Do Models Believe What They Say?","created":"2026-06-09","cited_by_count":4,"area":null},{"id":"2506.16975","title":"Latent Concept Disentanglement in Transformer-based Language Models","created":"2025-06-20","cited_by_count":3,"area":null},{"id":"2510.00319","title":"DecepChain: Inducing Deceptive Reasoning in Large Language Models","created":"2025-09-30","cited_by_count":3,"area":"lr"},{"id":"2602.00298","title":"Assessing Domain-Level Susceptibility to Emergent Misalignment from Narrow Finetuning","created":"2026-01-30","cited_by_count":3,"area":"safety"},{"id":"2602.10371","title":"Simple LLM Baselines are Competitive for Model Diffing","created":"2026-02-10","cited_by_count":3,"area":"mechinterp"},{"id":"2604.09544","title":"Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types","created":"2026-04-10","cited_by_count":3,"area":"safety"},{"id":"2605.31328","title":"Reinforcement Learning Can Amplify Emergent Misalignment from Harmless Rewards","created":"2026-05-29","cited_by_count":3,"area":"alignment"},{"id":"2606.30252","title":"Inoculation Adapters: Improved Selective Generalization of Capabilities with Fewer Surprising Backdoors","created":"2026-06-29","cited_by_count":3,"area":null},{"id":"2607.21356","title":"Emergent Misalignment Recruits a Pre-existing Persona Subspace","created":"2026-07-23","cited_by_count":3,"area":"alignment"},{"id":"2510.07775","title":"The Unintended Trade-off of AI Alignment:Balancing Hallucination Mitigation and Safety in LLMs","created":"2025-10-09","cited_by_count":2,"area":"safety"},{"id":"2606.03810","title":"Consistency Training Can Entrench Misalignment","created":"2026-06-02","cited_by_count":2,"area":null},{"id":"2606.06667","title":"The Piggyback Hypothesis of Generalization: Explaining and Mitigating Emergent Misalignment","created":"2026-06-04","cited_by_count":2,"area":null},{"id":"2606.08682","title":"Activation Steering Induces Emergent Misalignment: A More Comprehensive Evaluation","created":"2026-06-07","cited_by_count":2,"area":"mechinterp"},{"id":"2607.04432","title":"Covert Trait Propagation Is Representation Alignment: Mechanistic Evidence from Hidden-Channel Distillation","created":"2026-07-05","cited_by_count":2,"area":null},{"id":"2607.27617","title":"Hidden APIs in Language Models: Discovering Reusable Causal Interfaces from Forked Futures","created":"2026-07-30","cited_by_count":2,"area":null},{"id":"2601.04369","title":"Generalization to Political Beliefs from Fine-Tuning on Sports Team Preferences","created":"2026-01-07","cited_by_count":1,"area":null},{"id":"2603.12397","title":"Not Just the Destination, But the Journey: Reasoning Traces Causally Shape Generalization Behaviors","created":"2026-03-12","cited_by_count":1,"area":"lr"},{"id":"2603.25861","title":"Why Safety Probes Catch Liars But Miss Fanatics","created":"2026-03-26","cited_by_count":1,"area":null},{"id":"2604.10022","title":"Weird Generalization is Weirdly Brittle","created":"2026-04-11","cited_by_count":1,"area":null},{"id":"2604.16697","title":"Surgical Repair of Insecure Code Generation in LLMs","created":"2026-04-17","cited_by_count":1,"area":null},{"id":"2604.17663","title":"ATLAS: Constitution-Conditioned Latent Geometry and Redistribution Across Language Models and Neural Perturbation Data","created":"2026-04-19","cited_by_count":1,"area":null},{"id":"2604.28082","title":"Characterizing the Consistency of the Emergent Misalignment Persona","created":"2026-04-30","cited_by_count":1,"area":null},{"id":"2605.07462","title":"The Moltbook Files: A Harmless Slopocalypse or Humanity's Last Experiment","created":"2026-05-08","cited_by_count":1,"area":null},{"id":"2605.27676","title":"Unsupervised Identification and Removal of Spurious Correlations During Fine-Tuning","created":"2026-05-26","cited_by_count":1,"area":null},{"id":"2606.04413","title":"(Mis)generalization of Helpful-only Fine-tuning","created":"2026-06-03","cited_by_count":1,"area":null},{"id":"2606.05817","title":"Consistency Training Along the Transformer Stack","created":"2026-06-04","cited_by_count":1,"area":null},{"id":"2606.09068","title":"Emergent Misalignment Can Be Induced by Sycophancy and Reversed via Alignment Gating","created":"2026-06-08","cited_by_count":1,"area":null},{"id":"2606.20225","title":"Actionable Activation Directions for Detecting and Mitigating Emergent Misalignment Across Language Model Families","created":"2026-06-18","cited_by_count":1,"area":null},{"id":"2606.23700","title":"Self-Recognition Finetuning can Prevent and Reverse Emergent Misalignment","created":"2026-06-04","cited_by_count":1,"area":"safety"},{"id":"2607.26173","title":"Shared SFT Lessons Across Alignment, Model Organisms, and Toy Models","created":"2026-07-28","cited_by_count":1,"area":null},{"id":"2608.11025","title":"Data Attribution of Emergent Misalignment with Persona Features","created":"2026-08-11","cited_by_count":1,"area":"mechinterp"},{"id":"2609.01091","title":"Subliminal Learning as Trait-Direction Drift: A Mechanism and Targeted Control under SFT Distillation","created":"2026-09-01","cited_by_count":1,"area":"distill"},{"id":"2609.23935","title":"Measuring the Assistant's Harmlessness Preferences on the User Turn","created":"2026-09-20","cited_by_count":1,"area":null},{"id":"2609.25518","title":"Matryoshka attribution: Learning to attribute language model outputs to representations and weights","created":"2026-09-22","cited_by_count":1,"area":null},{"id":"2609.37914","title":"The Unequal Influence of Bad Advice: Using Training Data Attribution to Modulate Emergent Misalignment","created":"2026-09-29","cited_by_count":1,"area":null},{"id":"2601.03388","title":"Metaphors are a Source of Cross-Domain Misalignment of Large Reasoning Models","created":"2026-01-06","cited_by_count":0,"area":"lr"},{"id":"2602.02304","title":"Comparing Explanations is Not Enough, Explain the Change: New Standards are Needed to Explain Behavioral Shifts in Large Language Models","created":"2026-02-02","cited_by_count":0,"area":"alignment"},{"id":"2603.06816","title":"\"Dark Triad\" Model Organisms of Misalignment: Narrow Fine-Tuning Mirrors Human Antisocial Behavior","created":"2026-03-06","cited_by_count":0,"area":null},{"id":"2605.00994","title":"Most Current Model Organisms Are Leaky: Perplexity Differencing Often Reveals Finetuning Objectives","created":"2026-05-01","cited_by_count":0,"area":null},{"id":"2605.01130","title":"Iterative Finetuning is Mostly Idempotent","created":"2026-05-01","cited_by_count":0,"area":"alignment"},{"id":"2605.09773","title":"Exploitation Without Deception: Dark Triad Feature Steering Reveals Separable Antisocial Circuits in Language Models","created":"2026-05-10","cited_by_count":0,"area":"mechinterp"},{"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":"2606.07631","title":"Trait-space Monitoring for Emergent Misalignment During Supervised Finetuning","created":"2026-05-31","cited_by_count":0,"area":null},{"id":"2606.10747","title":"The Arbiter Agent: Continually Monitoring Multi-Agent Conversations to Detect Emergent Misalignment","created":"2026-06-09","cited_by_count":0,"area":"agents"},{"id":"2606.31591","title":"Evil Spectra: How Optimisers can Amplify or Suppress Emergent Misalignment","created":"2026-06-30","cited_by_count":0,"area":null},{"id":"2608.06123","title":"Poli-Bias: Understanding and Measuring Large Language Model Biases in International Political Conflicts","created":"2026-08-06","cited_by_count":0,"area":null},{"id":"2608.08212","title":"Harmful Content Is Not Enough: Continuation Framing Moderates In-Context Emergent Misalignment","created":"2026-08-08","cited_by_count":0,"area":"safety"},{"id":"2608.23476","title":"On the Threat Model of Weird Generalization and Emergent Misalignment","created":"2026-08-24","cited_by_count":0,"area":null},{"id":"2608.30619","title":"Hidden Threat in Synthetic Data: Covert Targeted Bias Injection through Benign Text","created":"2026-08-31","cited_by_count":0,"area":"synth"},{"id":"2609.15422","title":"Empirical Evaluation of Task-Based Permission Scoping Architecture for AI Agents","created":"2026-09-14","cited_by_count":0,"area":null},{"id":"2609.17865","title":"Do Frontier Models Seek Safety Evidence Before Acting?","created":"2026-09-15","cited_by_count":0,"area":null},{"id":"2609.20129","title":"Local Sparsity Enables Unsupervised LLM Safety Detection","created":"2026-09-17","cited_by_count":0,"area":"safety"},{"id":"2609.24504","title":"On Emergent Capabilities and Model Merging","created":"2026-09-21","cited_by_count":0,"area":null},{"id":"2609.34970","title":"See it, Say it, Sorted: Mechanistic Diagnosis and Parameter-Space Mitigation of Emergent Misalignment in LLMs","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.35291","title":"Narrow Multimodal Fine-Tuning Can Induce Emergent Misalignment","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2609.37624","title":"Correct, Don't Delete: Mitigating Emergent Misalignment with Corrective Supervision","created":"2026-09-29","cited_by_count":0,"area":null},{"id":"2609.38093","title":"Character Training for Risk-Averse Agents","created":"2026-09-29","cited_by_count":0,"area":null},{"id":"2609.38379","title":"Aligned Data Can Induce Misalignment via Context Confusion","created":"2026-09-29","cited_by_count":0,"area":null},{"id":"2610.00767","title":"Pre-training interventions, ex post facto: Grafting model beliefs across checkpoints","created":"2026-09-30","cited_by_count":0,"area":null},{"id":"2610.10203","title":"How to train your model organism","created":"2026-10-07","cited_by_count":0,"area":"safety"},{"id":"2610.12129","title":"An Investigation of Model Coherence: Narrow Finetunes Contradict Themselves Under Resampling","created":"2026-10-08","cited_by_count":0,"area":null}],"next_cursor":null}