Copertina

Lessons from Defending Gemini Against Indirect Prompt Injections

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Shi, Chongyang, Lin, Sharon, Song, Shuang, Hayes, Jamie, Shumailov, Ilia, Yona, Itay, Pluto, Juliette, Pappu, Aneesh, Choquette-Choo, Christopher A., Nasr, Milad, Sitawarin, Chawin, Gibson, Gena, Terzis, Andreas, Flynn, John "Four"
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Cryptography and Security Machine Learning
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Extracting alignment data in open models
di: Barbero, Federico, et al.
Pubblicazione: (2025)

The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections
di: Nasr, Milad, et al.
Pubblicazione: (2025)

Buffer Overflow in Mixture of Experts
di: Hayes, Jamie, et al.
Pubblicazione: (2024)

Measuring memorization in language models via probabilistic extraction
di: Hayes, Jamie, et al.
Pubblicazione: (2024)

Measuring memorization in RLHF for code completion
di: Pappu, Aneesh, et al.
Pubblicazione: (2024)

Soft Instruction De-escalation Defense
di: Walter, Nils Philipp, et al.
Pubblicazione: (2025)

Stealing User Prompts from Mixture of Experts
di: Yona, Itay, et al.
Pubblicazione: (2024)

StruQ: Defending Against Prompt Injection with Structured Queries
di: Chen, Sizhe, et al.
Pubblicazione: (2024)

Defending Against Prompt Injection With a Few DefensiveTokens
di: Chen, Sizhe, et al.
Pubblicazione: (2025)

Interpreting the Repeated Token Phenomenon in Large Language Models
di: Yona, Itay, et al.
Pubblicazione: (2025)

Cascading Adversarial Bias from Injection to Distillation in Language Models
di: Chaudhari, Harsh, et al.
Pubblicazione: (2025)

Defeating Prompt Injections by Design
di: Debenedetti, Edoardo, et al.
Pubblicazione: (2025)

Thought-Transfer: Indirect Targeted Poisoning Attacks on Chain-of-Thought Reasoning Models
di: Chaudhari, Harsh, et al.
Pubblicazione: (2026)

Positional Embedding-Aware Activations
di: Shah, Kathan, et al.
Pubblicazione: (2023)

Operationalizing Contextual Integrity in Privacy-Conscious Assistants
di: Ghalebikesabi, Sahra, et al.
Pubblicazione: (2024)

PubDef: Defending Against Transfer Attacks From Public Models
di: Sitawarin, Chawin, et al.
Pubblicazione: (2023)

Auditing Private Prediction
di: Chadha, Karan, et al.
Pubblicazione: (2024)

The Last Iterate Advantage: Empirical Auditing and Principled Heuristic Analysis of Differentially Private SGD
di: Steinke, Thomas, et al.
Pubblicazione: (2024)

Privacy Auditing of Large Language Models
di: Panda, Ashwinee, et al.
Pubblicazione: (2025)

UnUnlearning: Unlearning is not sufficient for content regulation in advanced generative AI
di: Shumailov, Ilia, et al.
Pubblicazione: (2024)

Large Language Models Can Verbatim Reproduce Long Malicious Sequences
di: Lin, Sharon, et al.
Pubblicazione: (2025)

OODRobustBench: a Benchmark and Large-Scale Analysis of Adversarial Robustness under Distribution Shift
di: Li, Lin, et al.
Pubblicazione: (2023)

PAL: Proxy-Guided Black-Box Attack on Large Language Models
di: Sitawarin, Chawin, et al.
Pubblicazione: (2024)

Jatmo: Prompt Injection Defense by Task-Specific Finetuning
di: Piet, Julien, et al.
Pubblicazione: (2023)

Defending Against Indirect Prompt Injection Attacks With Spotlighting
di: Hines, Keegan, et al.
Pubblicazione: (2024)

Defending against Indirect Prompt Injection by Instruction Detection
di: Wen, Tongyu, et al.
Pubblicazione: (2025)

Beyond Slow Signs in High-fidelity Model Extraction
di: Foerster, Hanna, et al.
Pubblicazione: (2024)

Mark My Words: Analyzing and Evaluating Language Model Watermarks
di: Piet, Julien, et al.
Pubblicazione: (2023)

Benchmarking and Defending Against Indirect Prompt Injection Attacks on Large Language Models
di: Yi, Jingwei, et al.
Pubblicazione: (2023)

Privacy Side Channels in Machine Learning Systems
di: Debenedetti, Edoardo, et al.
Pubblicazione: (2023)

LLMs unlock new paths to monetizing exploits
di: Carlini, Nicholas, et al.
Pubblicazione: (2025)

Inexact Unlearning Needs More Careful Evaluations to Avoid a False Sense of Privacy
di: Hayes, Jamie, et al.
Pubblicazione: (2024)

Attention is All You Need to Defend Against Indirect Prompt Injection Attacks in LLMs
di: Zhong, Yinan, et al.
Pubblicazione: (2025)

Exploring the limits of strong membership inference attacks on large language models
di: Hayes, Jamie, et al.
Pubblicazione: (2025)

The Task Shield: Enforcing Task Alignment to Defend Against Indirect Prompt Injection in LLM Agents
di: Jia, Feiran, et al.
Pubblicazione: (2024)

ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior
di: Lu, Weikai, et al.
Pubblicazione: (2025)

SEA: Shareable and Explainable Attribution for Query-based Black-box Attacks
di: Gao, Yue, et al.
Pubblicazione: (2023)

Machine Learning needs Better Randomness Standards: Randomised Smoothing and PRNG-based attacks
di: Dahiya, Pranav, et al.
Pubblicazione: (2023)

Fairness Feedback Loops: Training on Synthetic Data Amplifies Bias
di: Wyllie, Sierra, et al.
Pubblicazione: (2024)

The Landscape of Memorization in LLMs: Mechanisms, Measurement, and Mitigation
di: Xiong, Alexander, et al.
Pubblicazione: (2025)