No article text was recorded for this version.
instruction-following vulnerability (capability-security tradeoff)
Version 1 of 131 papers50 conceptsfrozen 2026-10-04
Papers in this version
- Blog: How Not to Detect Prompt Injections with an LLM
- Blog: How Trustworthy are Open-Source LLMs? An Assessment under Malicious Demonstrations Shows their Vulnerabilities
- Blog: Rapid Adoption, Hidden Risks: The Dual Impact of Large Language Model Customization
- Blog: DecodingTrust: A Comprehensive Assessment of Trustworthiness in GPT Models
- Blog: Follow My Instruction and Spill the Beans: Scalable Data Extraction from Retrieval-Augmented Generation Systems
- Blog: When "Competency" in Reasoning Opens the Door to Vulnerability: Jailbreaking LLMs via Novel Ciphers
- Blog: Evaluating the Instruction-Following Robustness of Large Language Models to Prompt Injection
- Blog: Efficient but Vulnerable: Benchmarking and Defending LLM Batch Prompting Attack
- Blog: Emerging Cyber Attack Risks of Medical AI Agents
- Blog: Mimicking the Familiar: Dynamic Command Generation for Information Theft Attacks in LLM Tool-Learning System
- Blog: Raccoon: Prompt Extraction Benchmark of LLM-Integrated Applications
- Blog: Ranking Manipulation for Conversational Search Engines
- Blog: Understanding and Evaluating Claw-like Agent Security Through a Computer-Systems Lens
- Blog: From Allies to Adversaries: Manipulating LLM Tool-Calling through Adversarial Injection
- Blog: An LLM can Fool Itself: A Prompt-Based Adversarial Attack
- Blog: Prompt Injection Attacks on LLM Generated Reviews of Scientific Publications
- Blog: Backdoor Attacks for In-Context Learning with Language Models
- Blog: Sell Me This Stock: Unsafe Recommendation Drift in LLM Agents
- Blog: Distract Large Language Models for Automatic Jailbreak Attack
- Blog: FLEX: A Benchmark for Evaluating Robustness of Fairness in Large Language Models
- Blog: Does AI Reviewer See the Full Picture? Attacking and Defending Multimodal Peer Review
- Blog: GRADEO: Towards Human-Like Evaluation for Text-to-Video Generation via Multi-Step Reasoning
- Blog: Adaptive Probe-based Steering for Robust LLM Jailbreaking
- Blog: Benchmarking adversarial robustness to bias elicitation in large language models: scalable automated assessment with LLM-as-a-judge
- Blog: ATLAS: Adaptive Topology-Level Attack Synthesis for Probing Multi-Agent Systems
- Blog: MCP-SafetyBench: A Benchmark for Safety Evaluation of Large Language Models with Real-World MCP Servers
- Blog: Black-Box Adversarial Attacks on LLM-Based Code Completion
- Blog: Benchmarking Knowledge-Extraction Attack and Defense on Retrieval-Augmented Generation
- Blog: Agent Security Bench (ASB): Formalizing and Benchmarking Attacks and Defenses in LLM-based Agents
- Blog: ASSERT: Automated Safety Scenario Red Teaming for Evaluating the Robustness of Large Language Models
- Blog: Subgraph Reconstruction Attacks on Graph RAG Deployments with Practical Defenses
History & citation
Current · Permanent link
- This direction split from The Inverse Scaling of Model Capability and Security (87% overlap)
Cite this direction
Latest version (a living page):
@misc{lacuna_instruction_following_vulnerability_capa_c8149951,
title = {instruction-following vulnerability (capability-security tradeoff)},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/instruction-following-vulnerability-capability-security-tradeoff/rdr_c814995179e845299cf0f2bd68b7ca16},
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_capa_c8149951_v1,
title = {instruction-following vulnerability (capability-security tradeoff)},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/instruction-following-vulnerability-capability-security-tradeoff/rdr_c814995179e845299cf0f2bd68b7ca16/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-21},
}