UNHSAILLab
Popular repositories Loading
-
working-memory-attack-on-llms
working-memory-attack-on-llms PublicWorking Memory Attack on LLMs
-
SentimentalLIAR
SentimentalLIAR PublicOur Sentimental LIAR dataset is a modified and further extended version of the LIAR extension introduced by Kirilin et al. In our dataset, the multi-class labeling of LIAR is converted to a binary …
-
Adversary-Engagement-Ontology
Adversary-Engagement-Ontology PublicThe adversary engagement ontology for expressing all things cyber denial, deception, and operational narratives.
-
medgemma-explainer
medgemma-explainer PublicForked from thedatasense/medgemma-explainer
This library generates relevancy maps that highlight which image regions and text tokens contribute to MedGemma's predictions. It implements the method from: Chefer, H., Gur, S., & Wolf, L. (2021).…
-
Phorcys-AutoPT-Framework
Phorcys-AutoPT-Framework PublicCapstone project for using Reinforcement Learning to conduct intelligent penetration tests.
Repositories
- MedMCQA-Robustness-Study Public
A comprehensive research framework for evaluating the robustness of large language models on medical multiple-choice question (MCQ) tasks. This project systematically tests how prompt configurations, option ordering, evidence context, and self-consistency voting impact model performance on medical knowledge benchmarks.
UNHSAILLab/MedMCQA-Robustness-Study’s past year of commit activity - medgemma-explainer Public Forked from thedatasense/medgemma-explainer
This library generates relevancy maps that highlight which image regions and text tokens contribute to MedGemma's predictions. It implements the method from: Chefer, H., Gur, S., & Wolf, L. (2021). Transformer Interpretability Beyond Attention Visualization. CVPR 2021.
UNHSAILLab/medgemma-explainer’s past year of commit activity - VSF-Med Public
We present VSF-Med, a Vulnerability Scoring Framework for medical Vision–Language Models (Vision LLMs), designed to systematically evaluate safety, reliability, and adversarial robustness in clinical imaging applications.
UNHSAILLab/VSF-Med’s past year of commit activity - TaCo Public
TaCo: Enhancing Cross-Lingual Transfer for Low-Resource Languages in LLMs through Translation-Assisted Chain-of-Thought Processes
UNHSAILLab/TaCo’s past year of commit activity - sandwich-attack Public
UNHSAILLab/sandwich-attack’s past year of commit activity
Top languages
Loading…
Most used topics
Loading…