Xirui Li

DrAttack. Official implementation of paper: DrAttack: Prompt Decomposition and Reconstruction Makes Powerful LLM Jailbreakers

ClawEnvKit. Open-source Environment toolkit of claw-like agents, support task/harness generation and evaluation

MOSSBench. An implementation for MLLM oversensitivity evaluation

Interpretation_DETR. A visual interpretation tool for Deformable DETR

Frankenstein. The official implementation for the paper "What does RL improve for Visual Reasoning? A Frankenstein-Style Analysis""

war-test. The official implementation for the paper "When AI Navigates the Fog of War"

attacks-on-LLMs. Awesome list for attacks on large language models.

RPN_with_Deformable_DETR. Python

Diffusion_Interpretation. Python

Agent-Mem. Python

MoltBook-Evaluation. The evaluation framework designed for AI agent-only society