Curated archive of jailbreak and prompt-injection experiments for AI model robustness research, red-teaming, and defensive evaluation.
VISION is a framework for robust and interpretable code vulnerability detection using counterfactual data augmentation. It leverages GNNs, LLM-generated counterfactuals, and graph-based explainability to mitigate spurious correlations and improve generalization on real-world vulnerabilities (CWE-20).
A Framework for Robust, Self-Recovering Tool-Using Language Model Agents — trained on 50K+ failure-annotated trajectories for fault-tolerant reasoning and recovery.
Next.js autonomous AI red-teaming, jailbreak robustness evaluation suite and multi-LLM comparative research interface.
A modular AI evaluation framework exploring mechanistic interpretability and RAG-based explainability. Built for EdTech grading and assessment tools (K-12, HigherEd, Lifelong), but adaptable to other high-risk compliance markets (e.g., HRTech).
Investigating the "Gradient Noise Paradox" in AI Safety: A study on the conflict between Differential Privacy (DP-SGD) and Adversarial Training. Uses a custom "Shadow Model" pipeline to synchronize Opacus with PGD attacks, demonstrating how privacy-preserving noise systematically degrades model robustness
To associate your repository with the ai-robustness topic, visit your repo's landing page and select "manage topics."