Tell me about yourself: LLMs are aware of their learned behaviors
Fuente:
arXiv
Saved in:
| Main Authors: | Betley, Jan, Bao, Xuchan, Soto, Martín, Sztyber-Betley, Anna, Chua, James, Evans, Owain |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Similar Items
Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs
by: Betley, Jan, et al.
Published: (2025)
by: Betley, Jan, et al.
Published: (2025)
Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs
by: Betley, Jan, et al.
Published: (2025)
by: Betley, Jan, et al.
Published: (2025)
Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers
by: Dubiński, Jan, et al.
Published: (2026)
by: Dubiński, Jan, et al.
Published: (2026)
Subliminal Learning: Language models transmit behavioral traits via hidden signals in data
by: Cloud, Alex, et al.
Published: (2025)
by: Cloud, Alex, et al.
Published: (2025)
Thought Crime: Backdoors and Emergent Misalignment in Reasoning Models
by: Chua, James, et al.
Published: (2025)
by: Chua, James, et al.
Published: (2025)
School of Reward Hacks: Hacking harmless tasks generalizes to misaligned behavior in LLMs
by: Taylor, Mia, et al.
Published: (2025)
by: Taylor, Mia, et al.
Published: (2025)
The Consciousness Cluster: Emergent preferences of Models that Claim to be Conscious
by: Chua, James, et al.
Published: (2026)
by: Chua, James, et al.
Published: (2026)
Jailbroken Frontier Models Retain Their Capabilities
by: Zhu, Daniel, et al.
Published: (2026)
by: Zhu, Daniel, et al.
Published: (2026)
Connecting the Dots: LLMs can Infer and Verbalize Latent Structure from Disparate Training Data
by: Treutlein, Johannes, et al.
Published: (2024)
by: Treutlein, Johannes, et al.
Published: (2024)
Signal in the Noise: Polysemantic Interference Transfers and Predicts Cross-Model Influence
by: Gong, Bofan, et al.
Published: (2025)
by: Gong, Bofan, et al.
Published: (2025)
SeedPrints: Fingerprints Can Even Tell Which Seed Your Large Language Model Was Trained From
by: Tong, Yao, et al.
Published: (2025)
by: Tong, Yao, et al.
Published: (2025)
Supporting Artifact Evaluation with LLMs: A Study with Published Security Research Papers
by: Heye, David, et al.
Published: (2026)
by: Heye, David, et al.
Published: (2026)
Me, Myself, and AI: The Situational Awareness Dataset (SAD) for LLMs
by: Laine, Rudolf, et al.
Published: (2024)
by: Laine, Rudolf, et al.
Published: (2024)
T-MAP: Red-Teaming LLM Agents with Trajectory-aware Evolutionary Search
by: Lee, Hyomin, et al.
Published: (2026)
by: Lee, Hyomin, et al.
Published: (2026)
Context Misleads LLMs: The Role of Context Filtering in Maintaining Safe Alignment of LLMs
by: Kim, Jinhwa, et al.
Published: (2025)
by: Kim, Jinhwa, et al.
Published: (2025)
RedAgent: Red Teaming Large Language Models with Context-aware Autonomous Language Agent
by: Xu, Huiyu, et al.
Published: (2024)
by: Xu, Huiyu, et al.
Published: (2024)
Defend LLMs Through Self-Consciousness
by: Huang, Boshi, et al.
Published: (2025)
by: Huang, Boshi, et al.
Published: (2025)
Protecting Your LLMs with Information Bottleneck
by: Liu, Zichuan, et al.
Published: (2024)
by: Liu, Zichuan, et al.
Published: (2024)
Bag of Tricks: Benchmarking of Jailbreak Attacks on LLMs
by: Xu, Zhao, et al.
Published: (2024)
by: Xu, Zhao, et al.
Published: (2024)
The Ethics of Interaction: Mitigating Security Threats in LLMs
by: Kumar, Ashutosh, et al.
Published: (2024)
by: Kumar, Ashutosh, et al.
Published: (2024)
Strategic Deflection: Defending LLMs from Logit Manipulation
by: Rachidy, Yassine, et al.
Published: (2025)
by: Rachidy, Yassine, et al.
Published: (2025)
PARASITE: Conditional System Prompt Poisoning to Hijack LLMs
by: Pham, Viet, et al.
Published: (2025)
by: Pham, Viet, et al.
Published: (2025)
Adversarial Tuning: Defending Against Jailbreak Attacks for LLMs
by: Liu, Fan, et al.
Published: (2024)
by: Liu, Fan, et al.
Published: (2024)
Running in CIRCLE? A Simple Benchmark for LLM Code Interpreter Security
by: Chua, Gabriel
Published: (2025)
by: Chua, Gabriel
Published: (2025)
Graph of Attacks: Improved Black-Box and Interpretable Jailbreaks for LLMs
by: Akbar-Tajari, Mohammad, et al.
Published: (2025)
by: Akbar-Tajari, Mohammad, et al.
Published: (2025)
Feint and Attack: Attention-Based Strategies for Jailbreaking and Protecting LLMs
by: Pu, Rui, et al.
Published: (2024)
by: Pu, Rui, et al.
Published: (2024)
SeqAR: Jailbreak LLMs with Sequential Auto-Generated Characters
by: Yang, Yan, et al.
Published: (2024)
by: Yang, Yan, et al.
Published: (2024)
Waterfall: Framework for Robust and Scalable Text Watermarking and Provenance for LLMs
by: Lau, Gregory Kang Ruey, et al.
Published: (2024)
by: Lau, Gregory Kang Ruey, et al.
Published: (2024)
bi-GRPO: Bidirectional Optimization for Jailbreak Backdoor Injection on LLMs
by: Ji, Wence, et al.
Published: (2025)
by: Ji, Wence, et al.
Published: (2025)
The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMs
by: Liu, Songyang, et al.
Published: (2025)
by: Liu, Songyang, et al.
Published: (2025)
Sandwich attack: Multi-language Mixture Adaptive Attack on LLMs
by: Upadhayay, Bibek, et al.
Published: (2024)
by: Upadhayay, Bibek, et al.
Published: (2024)
Less is More: Sparse Watermarking in LLMs with Enhanced Text Quality
by: Hoang, Duy C., et al.
Published: (2024)
by: Hoang, Duy C., et al.
Published: (2024)
Topology Matters: Measuring Memory Leakage in Multi-Agent LLMs
by: Liu, Jinbo, et al.
Published: (2025)
by: Liu, Jinbo, et al.
Published: (2025)
Rapid Optimization for Jailbreaking LLMs via Subconscious Exploitation and Echopraxia
by: Shen, Guangyu, et al.
Published: (2024)
by: Shen, Guangyu, et al.
Published: (2024)
Unlearning Backdoor Attacks for LLMs with Weak-to-Strong Knowledge Distillation
by: Zhao, Shuai, et al.
Published: (2024)
by: Zhao, Shuai, et al.
Published: (2024)
Attention Slipping: A Mechanistic Understanding of Jailbreak Attacks and Defenses in LLMs
by: Hu, Xiaomeng, et al.
Published: (2025)
by: Hu, Xiaomeng, et al.
Published: (2025)
What Features in Prompts Jailbreak LLMs? Investigating the Mechanisms Behind Attacks
by: Kirch, Nathalie, et al.
Published: (2024)
by: Kirch, Nathalie, et al.
Published: (2024)
ShieldLearner: A New Paradigm for Jailbreak Attack Defense in LLMs
by: Ni, Ziyi, et al.
Published: (2025)
by: Ni, Ziyi, et al.
Published: (2025)
Effective and Efficient Jailbreaks of Black-Box LLMs with Cross-Behavior Attacks
by: Gohil, Vasudev
Published: (2025)
by: Gohil, Vasudev
Published: (2025)
Stealthy Backdoor Attacks against LLMs Based on Natural Style Triggers
by: Wei, Jiali, et al.
Published: (2026)
by: Wei, Jiali, et al.
Published: (2026)
Similar Items
-
Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs
by: Betley, Jan, et al.
Published: (2025) -
Weird Generalization and Inductive Backdoors: New Ways to Corrupt LLMs
by: Betley, Jan, et al.
Published: (2025) -
Conditional misalignment: common interventions can hide emergent misalignment behind contextual triggers
by: Dubiński, Jan, et al.
Published: (2026) -
Subliminal Learning: Language models transmit behavioral traits via hidden signals in data
by: Cloud, Alex, et al.
Published: (2025) -
Thought Crime: Backdoors and Emergent Misalignment in Reasoning Models
by: Chua, James, et al.
Published: (2025)