Focused Papers, Delivered Simply ：）

PRISM: Robust VLM Alignment with Principled Reasoning for Integrated Safety in Multimodality

A toolbox for benchmarking Multimodal LLM Agents trustworthiness across truthfulness, controllability, safety and privacy dimensions through 34 interactive tasks

[ISMIR 2025] A curated list of vision-to-music generation: methods, datasets, evaluation and challenges.

[NeurIPS 2024] Official implementation for "AgentPoison: Red-teaming LLM Agents via Memory or Knowledge Base Backdoor Poisoning"

Official codebase for "STAIR: Improving Safety Alignment with Introspective Reasoning"

AISafetyLab: A comprehensive framework covering safety attack, defense, evaluation and paper list.

Codebase for PINNacle: A Comprehensive Benchmark of Physics-Informed Neural Networks for Solving PDEs.

Open source AI/ML capabilities for the FiftyOne ecosystem

DanceCamAnimator: Keyframe-Based Controllable 3D Dance Camera Synthesis. [ACMMM 2024] Official PyTorch implementation

[NTIRE2024] official code for "Towards Real-world Video Face Restoration: A New Benchmark"

Adversarial Distributional Training (NeurIPS 2020)

[NeurIPS 2023] ImageReward: Learning and Evaluating Human Preferences for Text-to-image Generation

A toolbox for benchmarking trustworthiness of multimodal large language models (MultiTrust, NeurIPS 2024 Track Datasets and Benchmarks)

[NeurIPS 2024 Spotlight ⭐️ & TPAMI 2025] Parameter-Inverted Image Pyramid Networks (PIIP)

[CVPR2024 Highlight] Official implementation for Transferable Visual Prompting. The paper "Exploring the Transferability of Visual Prompting for Multimodal Large Language Models" has been accepted …

[EMNLP 2023] Lion: Adversarial Distillation of Proprietary Large Language Models

Beyond the Imitation Game collaborative benchmark for measuring and extrapolating the capabilities of language models

TruthfulQA: Measuring How Models Imitate Human Falsehoods

Human preference data for "Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback"

An open platform for training, serving, and evaluating large language models. Release repo for Vicuna and Chatbot Arena.

Safe RLHF: Constrained Value Alignment via Safe Reinforcement Learning from Human Feedback

Refine high-quality datasets and visual AI models

Code and documentation to train Stanford's Alpaca models, and generate the data.