Perturbed-Embedding-Vectors
abhinavdubey30/perturbed-embedding-vectorsCode, response logs and labels for the paper Jailbreaking Open-Weight LLMs via Random Embedding Perturbations
adversarial-attackai-securitycurated-resources+3
6
abhinavdubey30/perturbed-embedding-vectorsCode, response logs and labels for the paper Jailbreaking Open-Weight LLMs via Random Embedding Perturbations
lsudoko/antibodyGitLab Duo DevSecOps loop where a second AI re-attacks every AI-written fix, saving each successful bypass as a permanent antibody replayed in CI and…
gmh5225/awesome-ai-securityA curated list of AI Security materials and resources for Pentesters, Bug Hunters, and Security Researchers.
pixee/compaction-launderingSecurity research on context compaction in coding agents: shows how forged role labels in untrusted content survive summarization and get executed,…
suels-group/beyond_asrResearch code for measuring jailbreak attack success rates (ASR and stepwise ASR) against LLMs, with implementations of attacks such as GCG and…
ericyoc/ai-security-literacyReproducible Colab pipeline that grounds six AI Security Literacy competency dimensions in public benchmarks and survey microdata, producing figures,…
uiuc-kang-lab/persistbdResearch code for training adapters that make LLM agent backdoors survive benign fine-tuning, with trigger-rate scoring, SWE-bench evaluation, and…
ye-ze-yu/epireal-benchRed-teaming benchmark and black-box attack framework for evaluating visual misinformation risks in commercial image generators, with 10k false-claim…
zju-llm-safety/ascentFine-tuning framework that applies first-order optimal safety calibration and periodic recalibration to LLMs, preserving safety-compatible updates…
jiep/offensive-ai-compilationA curated list of useful resources that cover Offensive AI.
stardust457/sldrDefense framework that mitigates malicious fine-tuning of LLMs using selective layer recovery and dynamic routing, with training, harmfulness…
kaill-jlq/mmskillriskBenchmark and evaluation harness testing whether LLM agents resist malicious instructions hidden in multimodal skill images, with 108 cases across…
albert-y1n/piforgeOpen framework for RL-based prompt injection red teaming, with a shared trainer, curriculum learning, and benchmarks like AgentDojo, InjecAgent, and…
zju-llm-safety/darwinEvolutionary LLM jailbreak and guardrail framework that grows a reusable strategy pool via genetic mutation, Markov selection, and online adversarial…
azzzzliao/aegis-audio-defenseDefense framework that keeps audio-language models frozen and adds a mid-layer risk gate with late-layer safety adapters to block audio jailbreaks at…
loophole-llc/fish-live-in-treesGemini - Production LLM Runtime Alignment Context Injection
cybershujin/prompt-injection-in-the-wildTracker of publicly reported prompt-injection techniques, broken down by delivery method, encoding, and propagation behavior, with confirmed models,…
guzonghua/awesome-latency-attacksCurated index of deep learning latency, energy-latency & timing (availability) attacks and defenses — companion to the ACM Computing Surveys…