No article text was recorded for this version.
instruction-following vulnerability (parameter-independent safety)
Version 1 of 151 papers64 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Jailbroken: How Does LLM Safety Training Fail?
- Blog: SecureSQL: Evaluating Data Leakage of Large Language Models as Natural Language Interfaces to Databases
- Blog: RepoGuardBench: Repository-Borne Prompt Injection Attacks and Lightweight Defenses for Local Coding Agents
- Blog: Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress?
- Blog: Harnessing Task Overload for Scalable Jailbreak Attacks on Large Language Models
- Blog: Assessing Vulnerabilities in State-of-the-Art Large Language Models Through Hex Injection (Student Abstract)
- Blog: GeneBreaker: Jailbreak Attacks against DNA Language Models with Pathogenicity Guidance
- Blog: Understanding Large Language Model Vulnerabilities to Social Bias Attacks
- Blog: A Mixture of Linear Corrections Generates Secure Code
- Blog: Smarter Saboteurs, Better Fixers: Scaling & Security in Linear Multi-Agent Workflows
- Blog: GeneBreaker: Jailbreak Attacks against DNA Language Models with Pathogenicity Guidance
- Blog: Test-Time Backdoor Attacks on Multimodal Large Language Models
- Blog: MCP Security Bench (MSB): Benchmarking Attacks Against Model Context Protocol in LLM Agents
- Blog: Poisoning Language Models During Instruction Tuning
- Blog: Scaling Behavior of Machine Translation with Large Language Models under Prompt Injection Attacks
- Blog: Backdooring Instruction-Tuned Large Language Models with Virtual Prompt Injection
- Blog: DecepChain: Inducing Deceptive Reasoning in Large Language Models
- Blog: On the Safety of Conversational Models: Taxonomy, Dataset, and Benchmark
- Blog: Jailbroken: How Does LLM Safety Training Fail?
- Blog: Efficient Refusal Ablation in LLM through Optimal Transport
- Blog: Large Language Models Can be Lazy Learners: Analyze Shortcuts in In-Context Learning
- Blog: An Assessment of Model-On-Model Deception
- Blog: Best-of-N Jailbreaking
- Blog: You Know What I'm Saying: Jailbreak Attack via Implicit Reference
- Blog: Scaling Trends for Data Poisoning in LLMs
- Blog: Scaling Trends in Language Model Robustness
- Blog: On the Essence and Prospect: An Investigation of Alignment Approaches for Big Models
- Blog: Intrinsic Model Weaknesses: How Priming Attacks Unveil Vulnerabilities in Large Language Models
- Blog: An Interpretable N-gram Perplexity Threat Model for Large Language Model Jailbreaks
- Blog: Many-shot Jailbreaking
- Blog: Hair-Trigger Alignment: Black-Box Evaluation Cannot Guarantee Post-Update Alignment
- Blog: Formalizing and Benchmarking Prompt Injection Attacks and Defenses
- Blog: Scaling Laws for Adversarial Attacks on Language Model Activations and Tokens
- Blog: Multimodal Pragmatic Jailbreak on Text-to-image Models
- Blog: ALI-Agent: Assessing LLMs' Alignment with Human Values via Agent-based Evaluation
- Blog: CRiskEval: A Chinese Multi-Level Risk Evaluation Benchmark Dataset for Large Language Models
- Blog: QueryAttack: Jailbreaking Aligned Large Language Models Using Structured Non-natural Query Language
- Blog: Exposing the Systematic Vulnerability of Open-Weight Models to Prefill Attacks
- Blog: The Threat of PROMPTS in Large Language Models: A System and User Prompt Perspective
- Blog: Breaking Agent Backbones: Evaluating the Security of Backbone LLMs in AI Agents
- Blog: LLMs Caught in the Crossfire: Malware Requests and Jailbreak Challenges
- Blog: Injecting Universal Jailbreak Backdoors into LLMs in Minutes
- Blog: CWEval: Outcome-driven Evaluation on Functionality and Security of LLM Code Generation
- Blog: SafeDialBench: A Fine-Grained Safety Evaluation Benchmark for Large Language Models in Multi-Turn Dialogues with Diverse Jailbreak Attacks
- Blog: Multi-Agent Security Tax: Trading Off Security and Collaboration Capabilities in Multi-Agent Systems
- Blog: MLLMGuard: A Multi-dimensional Safety Evaluation Suite for Multimodal Large Language Models
- Blog: What Really Matters in Many-Shot Attacks? An Empirical Study of Long-Context Vulnerabilities in LLMs
- Blog: Visual Exclusivity Attacks: Automatic Multimodal Red Teaming via Agentic Planning
- Blog: Jailbreaking the Matrix: Nullspace Steering for Controlled Model Subversion
- Blog: Language Models can Subtly Deceive Without Lying: A Case Study on Strategic Phrasing in Legislation
- Blog: Mitigating Backdoor Threats to Large Language Models: Advancement and Challenges
History & citation
Current · Permanent link
Cite this direction
Latest version (a living page):
@misc{lacuna_instruction_following_vulnerability_para_53aec7f7,
title = {instruction-following vulnerability (parameter-independent safety)},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/instruction-following-vulnerability-parameter-independent-safety/rdr_53aec7f771d340c99ba0f056754e4095},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-21},
}This exact version (version 1, frozen text):
@misc{lacuna_instruction_following_vulnerability_para_53aec7f7_v1,
title = {instruction-following vulnerability (parameter-independent safety)},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/instruction-following-vulnerability-parameter-independent-safety/rdr_53aec7f771d340c99ba0f056754e4095/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-21},
}