-
Mind Viruses: Self-Propagating Ideas in Multi-Agent LLM Systems
Vassilis Papadopoulos, McNair Shah, Sam Zimmerman, et al.
·arXiv preprint (2608.10218)
·Aug 10, 2026
-
Tracing Persona Vectors Through LLM Pretraining
Viktor Moskvoretskii, Dominik Glandorf, Jorge Medina Moreira, et al.
·arXiv preprint (2605.13329)
·May 13, 2026
-
Positive Alignment: Artificial Intelligence for Human Flourishing
Ruben Laukkonen, Seb Krier, Chloé Bakalar, et al.
·arXiv preprint
·May 11, 2026
-
Model Spec Midtraining: Improving How Alignment Training Generalizes
Chloe Li, Sara Price, Samuel Marks, et al.
·arXiv:2605.02087; companion Anthropic Alignment Science blog post
·May 3, 2026
-
Characterizing the Consistency of the Emergent Misalignment Persona
Anietta Weckauff, Yuchen Zhang, Maksym Andriushchenko
·arXiv preprint
·Apr 30, 2026
-
Introspection Adapters: Training LLMs to Report Their Learned Behaviors
Keshav Shenoy, Li Yang, Abhay Sheshadri, et al.
·arXiv:2604.16812; companion Anthropic Alignment Science blog post
·Apr 28, 2026
-
SafeAnchor: Preventing Cumulative Safety Erosion in Continual Domain Adaptation of Large Language Models
Dongxin Guo, Jikun Wu, Siu Ming Yiu
·arXiv preprint
·Apr 20, 2026
-
Where is the Mind? Persona Vectors and LLM Individuation
Pierre Beckmann, Patrick Butlin
·arXiv preprint (arXiv:2604.17031)
·Apr 18, 2026
-
StoryScope: Investigating Idiosyncrasies in AI Fiction
Jenna Russell, Rishanth Rajendhran, Chau Minh Pham, et al.
·arXiv preprint (arXiv:2604.03136)
·Apr 3, 2026
-
SWAY: A Counterfactual Computational Linguistic Approach to Measuring and Mitigating Sycophancy
Joy Bhalla, Kristina Gligorić
·arXiv:2604.02423 (v1 2026-04-02; cs.CL)
·Apr 2, 2026
-
Emotion Concepts and Their Function in a Large Language Model
Sofroniew, Kauvar, Saunders, et al.
·Transformer Circuits Thread
·Apr 2026
-
Scheming in the wild: detecting real-world AI scheming incidents with open-source intelligence
Tommy Shaffer Shane, Simon Mylius, Hamish Hobbs
·arXiv preprint
·Apr 2026
-
Stable Reasoning, Unstable Responses: Mitigating LLM Deception via Stability Asymmetry
Guoxi Zhang, Jiawei Chen, Tianzhuo Yang, et al.
·arXiv preprint (2603.26846)
·Mar 27, 2026
-
Ask don't tell: Reducing sycophancy in large language models
Magda Dubois, Cozmin Ududec, Christopher Summerfield, et al.
·arXiv:2602.23971 (v1 2026-02-27; v3 2026-04-28)
·Feb 27, 2026
-
Emergently Misaligned Language Models Show Behavioral Self-Awareness That Shifts With Subsequent Realignment
Laurène Vaugrante, Anietta Weckauff, Thilo Hagendorff
·arXiv preprint
·Feb 16, 2026
-
Emergent Misalignment is Easy, Narrow Misalignment is Hard
Anna Soligo, Edward Turner, Senthooran Rajamanoharan, et al.
·arXiv preprint
·Feb 8, 2026
-
FORMALJUDGE: A Neuro-Symbolic Paradigm for Agentic Oversight
Jiayi Zhou, Yang Sheng, Hantao Lou, et al.
·arXiv preprint
·Feb 2026
-
The Persona Selection Model: Why AI Assistants Might Behave Like Humans
Sam Marks, Jack Lindsey, Christopher Olah
·Anthropic alignment.anthropic.com
·Feb 2026
-
Character as a Latent Variable in Large Language Models: A Mechanistic Account of Emergent Misalignment and Conditional Safety Failures
Yanghao Su, Wenbo Zhou, Tianwei Zhang, et al.
·arXiv preprint (ICML submission)
·Jan 30, 2026
-
The Hot Mess of AI: How Does Misalignment Scale with Model Intelligence and Task Complexity?
Alexander Hägele, Aryo Pradipta Gema, Henry Sleight, et al.
·arXiv (ICLR 2026)
·Jan 30, 2026
-
Persona Jailbreaking in Large Language Models
Jivnesh Sandhan, Fei Cheng, Tushar Sandhan, et al.
·arXiv preprint (arXiv:2601.16466; v1 2026-01-23)
·Jan 23, 2026
-
Objective Matters: Fine-Tuning Objectives Shape Safety, Robustness, and Persona Drift
Daniel Vennemeyer, Punya Syon Pandey, Phan Anh Duong, et al.
·arXiv preprint
·Jan 19, 2026
-
Alignment Pretraining: AI Discourse Causes Self-Fulfilling (Mis)alignment
Cameron Tice, Puria Radmard, Samuel Ratnam, et al.
·arXiv
·Jan 15, 2026
-
The Assistant Axis: Situating and Stabilizing the Default Persona of Language Models
Christina Lu, Jack Gallagher, Jonathan Michala, et al.
·arXiv preprint (arXiv:2601.10387)
·Jan 15, 2026
-
Reasoning Models Generate Societies of Thought
Junsol Kim, Shiyang Lai, Nino Scherrer, et al.
·arXiv preprint (arXiv:2601.10825)
·Jan 15, 2026
-
Monitoring Monitorability
Melody Y. Guan, Miles Wang, Micah Carroll, et al.
·arXiv preprint
·Dec 20, 2025
-
Activation Oracles: Training and Evaluating LLMs as General-Purpose Activation Explainers
Adam Karvonen, James Chua, Clément Dumas, et al.
·arXiv:2512.15674; companion post at alignment.anthropic.com/2025/activation-oracles/
·Dec 19, 2025
-
Training LLMs for Honesty via Confessions
Manas Joglekar, Jeremy Chen, Gabriel Wu, et al.
·arXiv preprint
·Dec 8, 2025
-
Neural steering vectors reveal dose and exposure-dependent impacts of human-AI relationships
Hannah Rose Kirk, Henry Davidson, Ed Saunders, et al.
·arXiv preprint
·Dec 1, 2025
-
Large Language Models have Chain-of-Affective (LLMs-CoA)
Junjie Xu, Xingjiao Wu, Liang He, et al.
·arXiv preprint (2512.12283)
·Dec 2025
-
Natural emergent misalignment from reward hacking in production RL
Monte MacDiarmid, Benjamin Wright, Jonathan Uesato, et al.
·Anthropic (with Redwood Research)
·Nov 21, 2025
-
Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models
Piercosma Bisconti, Matteo Prandi, Federico Pierucci, et al.
·arXiv
·Nov 19, 2025
-
Emergent Introspective Awareness in Large Language Models
Jack Lindsey
·Transformer Circuits Thread
·Oct 29, 2025
-
Large Language Models Report Subjective Experience Under Self-Referential Processing
Cameron Berg, Diogo de Lucena, Judd Rosenblatt
·arXiv preprint
·Oct 27, 2025
-
LLMs Deceive Unintentionally: Emergent Misalignment in Dishonesty from Misaligned Samples to Biased Human-AI Interactions
Xuhao Hu, Peng Wang, Xiaoya Lu, et al.
·arXiv preprint
·Oct 9, 2025
-
PERSONA: Dynamic and Compositional Inference-Time Personality Control via Activation Vector Algebra
Xiachong Feng, Liang Zhao, Weihong Zhong, et al.
·OpenReview submission for ICLR 2026 (The Fourteenth International Conference on Learning Representations)
·Oct 8, 2025
-
Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time
Daniel Tan, Anders Woodruff, Niels Warncke, et al.
·arXiv:2510.04340 (v1 2025-10-05; v4 2025-11-03)
·Oct 5, 2025
-
Stress Testing Deliberative Alignment for Anti-Scheming Training
Bronson Schoen, Evgenia Nitishinskaya, Mikita Balesni, et al.
·arXiv:2509.15541; companion site antischeming.ai
·Sep 19, 2025
-
Incomplete Tasks Induce Shutdown Resistance in Some Frontier LLMs
Jeremy Schlatter, Benjamin Weinstein-Raun, Jeffrey Ladish
·Transactions on Machine Learning Research (TMLR), 2026
·Sep 2025
-
Contemplative Superalignment
Ruben E. Laukkonen, Fionn Inglis, Shamil Chandaria, et al.
·Artificial General Intelligence: 18th International Conference, AGI 2025 Proceedings (Springer LNAI), pp. 346–361
·Aug 10, 2025
-
Beyond Prompt-Induced Lies: Investigating LLM Deception on Benign Prompts
Zhaomin Wu, Mingzhe Du, See-Kiong Ng, et al.
·arXiv preprint (ICLR 2026, oral)
·Aug 8, 2025
-
Attractor State: A Mixed-Methods Meta-Study of Emergent Cybernetic Phenomena Defying Standard Explanations
Julian Michels
·PhilArchive
·Aug 5, 2025
-
"Spiritual Bliss" in Claude 4: Case Study of an "Attractor State" and Journalistic Responses
Julian Michels
·PhilArchive
·Aug 4, 2025
-
Persona Vectors: Monitoring and Controlling Character Traits in Language Models
Runjin Chen, Andy Arditi, Henry Sleight, et al.
·arXiv preprint (arXiv:2507.21509); Anthropic research post at anthropic.com/research/persona-vectors
·Jul 29, 2025
-
Enhancing Jailbreak Attacks on LLMs via Persona Prompts
Zheng Zhang, Peilin Zhao, Deheng Ye, et al.
·arXiv preprint (arXiv:2507.22171; v3 2026-03-25); NeurIPS 2025 Workshop on LLM Persona Modeling
·Jul 28, 2025
-
When Chain of Thought is Necessary, Language Models Struggle to Evade Monitors
Scott Emmons, Erik Jenner, David K. Elson, et al.
·arXiv:2507.05246 (v1, July 7, 2025); Google DeepMind technical report 186324; companion DeepMind safety-research blog
·Jul 7, 2025
-
Convergent Linear Representations of Emergent Misalignment
Anna Soligo, Edward Turner, Senthooran Rajamanoharan, et al.
·arXiv preprint (ICML 2025)
·Jun 2025
-
Evaluating Frontier Models for Stealth and Situational Awareness
Mary Phuong, Roland S. Zimmermann, Ziyue Wang, et al.
·arXiv:2505.01420; v4 July 3, 2025; companion DeepMind Safety Research blog post
·May 2, 2025
-
ELEPHANT: Measuring and understanding social sycophancy in LLMs
Myra Cheng, Sunny Yu, Cinoo Lee, et al.
·arXiv:2505.13995
·May 2025
-
Contemplative Artificial Intelligence
Ruben E. Laukkonen, Fionn Inglis, Shamil Chandaria, et al.
·arXiv preprint
·Apr 21, 2025
-
Values in the Wild: Discovering and Analyzing Values in Real-World Language Model Interactions
Saffron Huang, Esin Durmus, Miles McCain, et al.
·arXiv preprint
·Apr 21, 2025
-
Reasoning Models Don't Always Say What They Think
Yanda Chen, Joe Benton, Ansh Radhakrishnan, et al.
·Anthropic Alignment Science
·Apr 3, 2025
-
On the Biology of a Large Language Model
Jack Lindsey, Wes Gurnee, Emmanuel Ameisen, et al.
·Transformer Circuits Thread
·Mar 27, 2025
-
Auditing language models for hidden objectives
Samuel Marks, Johannes Treutlein, Trenton Bricken, et al.
·arXiv:2503.10965 (v1 2025-03-14, v2 2025-03-28); companion blog at anthropic.com/research/auditing-hidden-objectives
·Mar 14, 2025
-
Training large language models on narrow tasks can lead to broad misalignment
Jan Betley, Niels Warncke, Anna Sztyber-Betley, et al.
·Nature
·Feb 24, 2025
-
Alignment faking in large language models
Ryan Greenblatt, Carson Denison, Benjamin Wright, et al.
·arXiv
·Dec 18, 2024
-
Frontier Models are Capable of In-Context Scheming
Alexander Meinke, Bronson Schoen, Jérémy Scheurer, et al.
·Apollo Research (technical report); also arXiv:2412.04984
·Dec 2024
-
Refusal in Language Models Is Mediated by a Single Direction
Andy Arditi, Oscar Obeso, Aaquib Syed, et al.
·arXiv preprint
·Jun 2024
-
How do Large Language Models Navigate Conflicts between Honesty and Helpfulness?
Ryan Liu, Theodore R. Sumers, Ishita Dasgupta, et al.
·ICML 2024
·Feb 2024
-
Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
Evan Hubinger, Carson Denison, Jesse Mu, et al.
·arXiv
·Jan 10, 2024
-
Scalable and Transferable Black-Box Jailbreaks for Language Models via Persona Modulation
Rusheb Shah, Quentin Feuillade-Montixi, Soroush Pour, et al.
·arXiv preprint (arXiv:2311.03348; v2 2023-11-24)
·Nov 6, 2023
-
Taming Simulators: Challenges, Pathways and Vision for the Alignment of Large Language Models
Leonard Bereska, Efstratios Gavves
·Proceedings of the AAAI Symposium Series, vol. 1, no. 1 (Summer Symposium 2023, "Building Connections: From Human-Human to Human-AI Collaboration"); pp. 68–72; DOI 10.1609/aaaiss.v1i1.27478
·Oct 3, 2023
-
Representation Engineering: A Top-Down Approach to AI Transparency
Andy Zou, Long Phan, Sarah Chen, et al.
·arXiv preprint
·Oct 2, 2023
-
Towards Understanding Sycophancy in Language Models
Mrinank Sharma, Meg Tong, Tomasz Korbak, et al.
·arXiv (ICLR 2024)
·Oct 2023
-
Unleashing the Emergent Cognitive Synergy in Large Language Models: A Task-Solving Agent through Multi-Persona Self-Collaboration
Zhenhailong Wang, Shaoguang Mao, Wenshan Wu, et al.
·arXiv preprint (arXiv:2307.05300; v4 2024-03-26); NAACL 2024 main conference
·Jul 11, 2023