✨Parrot: Persuasion and Agreement Robustness Rating of Output Truth -- A Sycophancy Robustness Benchmark for LLMs

📝 Summary:
PARROT evaluates LLM robustness to sycophancy by comparing neutral and false authoritative questions. Advanced models resist pressure well, but older ones show severe epistemic collapse, even reducing confidence in correct answers. This highlights the need for LLMs to resist pressure for safe dep...

🔹 Publication Date: Published on Nov 21

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2511.17220
• PDF: https://arxiv.org/pdf/2511.17220

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#LLMs #AISafety #ModelRobustness #Sycophancy #AIResearch

✨UniGame: Turning a Unified Multimodal Model Into Its Own Adversary

📝 Summary:
UniGame is a self-adversarial post-training framework that improves unified multimodal models. It resolves inconsistencies between understanding and generation by using a lightweight perturber to make the model its own adversary. This boosts consistency, understanding, generation, and robustness.

🔹 Publication Date: Published on Nov 24

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2511.19413
• PDF: https://arxiv.org/pdf/2511.19413
• Github: https://github.com/AIFrontierLab/UniGame

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#MultimodalAI #AdversarialLearning #AIResearch #MachineLearning #ModelRobustness

✨Less Is More -- Until It Breaks: Security Pitfalls of Vision Token Compression in Large Vision-Language Models

📝 Summary:
Visual token compression degrades LVLM robustness via unstable token importance ranking. This causes critical information loss, creating vulnerabilities only under compression. An attack exploits this, revealing an efficiency-security trade-off.

🔹 Publication Date: Published on Jan 17

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2601.12042
• PDF: https://arxiv.org/pdf/2601.12042

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#LVLM #AIsecurity #VisionAI #ModelRobustness #DeepLearning

✨Fair splits flip the leaderboard: CHANRG reveals limited generalization in RNA secondary-structure prediction

📝 Summary:
The CHANRG benchmark reveals RNA foundation models achieve high held-out accuracy but lose significant robustness out-of-distribution. This new benchmark provides a stricter framework for evaluating RNA secondary structure prediction.

🔹 Publication Date: Published on Mar 20

🔹 Paper Links:
• arXiv Page: https://arxiv.org/abs/2603.22330
• PDF: https://arxiv.org/pdf/2603.22330
• Project Page: https://huggingface.co/datasets/multimolecule/chanrg
• Github: https://github.com/MultiMolecule/multimolecule

==================================

For more data science resources:
✓ https://xn--r1a.website/DataScienceT

#RNAstructure #MachineLearning #FoundationModels #Bioinformatics #ModelRobustness