Copertina

Emergent Misalignment is Easy, Narrow Misalignment is Hard

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Soligo, Anna, Turner, Edward, Rajamanoharan, Senthooran, Nanda, Neel
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Artificial Intelligence Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Convergent Linear Representations of Emergent Misalignment
di: Soligo, Anna, et al.
Pubblicazione: (2025)

Model Organisms for Emergent Misalignment
di: Turner, Edward, et al.
Pubblicazione: (2025)

Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models
di: Ferrando, Javier, et al.
Pubblicazione: (2024)

Thought Branches: Interpreting LLM Reasoning Requires Resampling
di: Macar, Uzay, et al.
Pubblicazione: (2025)

How Well Do Models Follow Their Constitutions?
di: Jakkli, Arya, et al.
Pubblicazione: (2026)

Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning
di: Casademunt, Helena, et al.
Pubblicazione: (2025)

Chain-of-Thought Reasoning In The Wild Is Not Always Faithful
di: Arcuschin, Iván, et al.
Pubblicazione: (2025)

Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs
di: Betley, Jan, et al.
Pubblicazione: (2025)

Re-Emergent Misalignment: How Narrow Fine-Tuning Erodes Safety Alignment in LLMs
di: Giordani, Jeremiah
Pubblicazione: (2025)

Semantic Containment as a Fundamental Property of Emergent Misalignment
di: Saxena, Rohan
Pubblicazione: (2026)

LLMs Deceive Unintentionally: Emergent Misalignment in Dishonesty from Misaligned Samples to Biased Human-AI Interactions
di: Hu, Xuhao, et al.
Pubblicazione: (2025)

Persona-Model Collapse in Emergent Misalignment
di: Costa, Davi Bastos, et al.
Pubblicazione: (2026)

Thought Crime: Backdoors and Emergent Misalignment in Reasoning Models
di: Chua, James, et al.
Pubblicazione: (2025)

Subliminal Learning Is Steering Vector Distillation
di: Blank, Camila, et al.
Pubblicazione: (2026)

Intrinsic Guardrails: How Semantic Geometry of Personality Interacts with Emergent Misalignment in LLMs
di: Aneja, Krishak, et al.
Pubblicazione: (2026)

Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
di: Lieberum, Tom, et al.
Pubblicazione: (2024)

Emergent and Subliminal Misalignment Through the Lens of Data-Mediated Transfer
di: Askin, Baris, et al.
Pubblicazione: (2026)

Are Sparse Autoencoders Useful? A Case Study in Sparse Probing
di: Kantamneni, Subhash, et al.
Pubblicazione: (2025)

Eliciting and Analyzing Emergent Misalignment in State-of-the-Art Large Language Models
di: Panpatil, Siddhant, et al.
Pubblicazione: (2025)

The Devil in the Details: Emergent Misalignment, Format and Coherence in Open-Weights LLMs
di: Dickson, Craig
Pubblicazione: (2025)

"Dark Triad" Model Organisms of Misalignment: Narrow Fine-Tuning Mirrors Human Antisocial Behavior
di: Lulla, Roshni, et al.
Pubblicazione: (2026)

Simple Mechanistic Explanations for Out-Of-Context Reasoning
di: Wang, Atticus, et al.
Pubblicazione: (2025)

Mitigating Misalignment Contagion by Steering with Implicit Traits
di: Chang, Maria, et al.
Pubblicazione: (2026)

Narrow Finetuning Leaves Clearly Readable Traces in Activation Differences
di: Minder, Julian, et al.
Pubblicazione: (2025)

Emergent misalignment as prompt sensitivity: A research note
di: Wyse, Tim, et al.
Pubblicazione: (2025)

Preemptive Detection and Correction of Misaligned Actions in LLM Agents
di: Fang, Haishuo, et al.
Pubblicazione: (2024)

Rethinking DPO: The Role of Rejected Responses in Preference Misalignment
di: Cho, Jay Hyeon, et al.
Pubblicazione: (2025)

Misaligned by Reward: Socially Undesirable Preferences in LLMs
di: Ghazaryan, Gayane, et al.
Pubblicazione: (2026)

Assessing Domain-Level Susceptibility to Emergent Misalignment from Narrow Finetuning
di: Mishra, Abhishek, et al.
Pubblicazione: (2026)

An Investigation into Value Misalignment in LLM-Generated Texts for Cultural Heritage
di: Bu, Fan, et al.
Pubblicazione: (2025)

RLHS: Mitigating Misalignment in RLHF with Hindsight Simulation
di: Liang, Kaiqu, et al.
Pubblicazione: (2025)

Dense SAE Latents Are Features, Not Bugs
di: Sun, Xiaoqing, et al.
Pubblicazione: (2025)

Residual Connections and the Causal Shift: Uncovering a Structural Misalignment in Transformers
di: Lys, Jonathan, et al.
Pubblicazione: (2026)

Bridging Draft Policy Misalignment: Group Tree Optimization for Speculative Decoding
di: Hu, Shijing, et al.
Pubblicazione: (2025)

Lost in Translation: Latent Concept Misalignment in Text-to-Image Diffusion Models
di: Zhao, Juntu, et al.
Pubblicazione: (2024)

Character as a Latent Variable in Large Language Models: A Mechanistic Account of Emergent Misalignment and Conditional Safety Failures
di: Su, Yanghao, et al.
Pubblicazione: (2026)

From Narrow Unlearning to Emergent Misalignment: Causes, Consequences, and Containment in LLMs
di: Mushtaq, Erum, et al.
Pubblicazione: (2025)

Explorations of Self-Repair in Language Models
di: Rushing, Cody, et al.
Pubblicazione: (2024)

Towards Best Practices of Activation Patching in Language Models: Metrics and Methods
di: Zhang, Fred, et al.
Pubblicazione: (2023)

Epistemic Traps: Rational Misalignment Driven by Model Misspecification
di: Xu, Xingcheng, et al.
Pubblicazione: (2026)