Cover Image

SafetyNet: Detecting Harmful Outputs in LLMs by Modeling and Monitoring Deceptive Behaviors

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Chaudhary, Maheep, Barez, Fazl
Format:	Preprint
Published:	2025
Subjects:	Artificial Intelligence Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Same Question, Different Words: A Latent Adversarial Framework for Prompt Robustness
by: Fu, Tingchen, et al.
Published: (2025)

In-Context Environments Induce Evaluation-Awareness in Language Models
by: Chaudhary, Maheep
Published: (2026)

Towards Interpretable Sequence Continuation: Analyzing Shared Circuits in Large Language Models
by: Lan, Michael, et al.
Published: (2023)

Interpreting Context Look-ups in Transformers: Investigating Attention-MLP Interactions
by: Neo, Clement, et al.
Published: (2024)

Beyond Linear Steering: Unified Multi-Attribute Control for Language Models
by: Oozeer, Narmeen, et al.
Published: (2025)

Scaling sparse feature circuit finding for in-context learning
by: Kharlapenko, Dmitrii, et al.
Published: (2025)

Quantifying Feature Space Universality Across Large Language Models via Sparse Autoencoders
by: Lan, Michael, et al.
Published: (2024)

MANATEE: Inference-Time Lightweight Diffusion Based Safety Defense for LLMs
by: Kan, Chun Yan Ryan, et al.
Published: (2026)

Alignment-Constrained Dynamic Pruning for LLMs: Identifying and Preserving Alignment-Critical Circuits
by: Patel, Dev, et al.
Published: (2025)

Weight space Detection of Backdoors in LoRA Adapters
by: Merenciano, David Puertolas, et al.
Published: (2026)

Understanding Addition in Transformers
by: Quirke, Philip, et al.
Published: (2023)

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training
by: Hubinger, Evan, et al.
Published: (2024)

Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs
by: Sheshadri, Abhay, et al.
Published: (2024)

DeceptionBench: A Comprehensive Benchmark for AI Deception Behaviors in Real-world Scenarios
by: Huang, Yao, et al.
Published: (2025)

Evaluating Open-Source Sparse Autoencoders on Disentangling Factual Knowledge in GPT-2 Small
by: Chaudhary, Maheep, et al.
Published: (2024)

Optimizing Chain-of-Thought Confidence via Topological and Dirichlet Risk Analysis
by: More, Abhishek, et al.
Published: (2025)

VAL-Bench: Belief Consistency as a measure for Value Alignment in Language Models
by: Gupta, Aman, et al.
Published: (2025)

Mechanistic origins of catastrophic forgetting: why RL preserves circuits better than SFT?
by: Nunez, Jeanmely Rojas, et al.
Published: (2026)

Best-of-N Jailbreaking
by: Hughes, John, et al.
Published: (2024)

Understanding Addition and Subtraction in Transformers
by: Quirke, Philip, et al.
Published: (2024)

Bypassing the Safety Training of Open-Source LLMs with Priming Attacks
by: Vega, Jason, et al.
Published: (2023)

Certifying Knowledge Comprehension in LLMs
by: Chaudhary, Isha, et al.
Published: (2024)

Deception Abilities Emerged in Large Language Models
by: Hagendorff, Thilo
Published: (2023)

Too Big to Fool: Resisting Deception in Language Models
by: Samsami, Mohammad Reza, et al.
Published: (2024)

Jailbreak Defense in a Narrow Domain: Limitations of Existing Methods and a New Transcript-Classifier Approach
by: Wang, Tony T., et al.
Published: (2024)

When Thinking LLMs Lie: Unveiling the Strategic Deception in Representations of Reasoning Models
by: Wang, Kai, et al.
Published: (2025)

Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content
by: Stepanov, Ihor, et al.
Published: (2026)

Detecting and Preventing Harmful Behaviors in AI Companions: Development and Evaluation of the SHIELD Supervisory System
by: Ben-Zion, Ziv, et al.
Published: (2025)

Towards Understanding Subliminal Learning: When and How Hidden Biases Transfer
by: Schrodi, Simon, et al.
Published: (2025)

Visualizing Neural Network Imagination
by: Wichers, Nevan, et al.
Published: (2024)

Unintended Harms of Value-Aligned LLMs: Psychological and Empirical Insights
by: Choi, Sooyung, et al.
Published: (2025)

LongSafety: Enhance Safety for Long-Context LLMs
by: Huang, Mianqiu, et al.
Published: (2024)

Eliciting Harmful Capabilities by Fine-Tuning On Safeguarded Outputs
by: Kaunismaa, Jackson, et al.
Published: (2026)

AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents
by: Andriushchenko, Maksym, et al.
Published: (2024)

Toxicity Detection Should Measure Contextual Harm, Not Text-Intrinsic Badness
by: Berezin, Sergei, et al.
Published: (2025)

SocialHarmBench: Revealing LLM Vulnerabilities to Socially Harmful Requests
by: Pandey, Punya Syon, et al.
Published: (2025)

Hydra: A Modular Architecture for Efficient Long-Context Reasoning
by: Chaudhary, Siddharth, et al.
Published: (2025)

Deception Detection from Linguistic and Physiological Data Streams Using Bimodal Convolutional Neural Networks
by: Li, Panfeng, et al.
Published: (2023)

Speak Easy: Eliciting Harmful Jailbreaks from LLMs with Simple Interactions
by: Chan, Yik Siu, et al.
Published: (2025)

SALT: Steering Activations towards Leakage-free Thinking in Chain of Thought
by: Batra, Shourya, et al.
Published: (2025)