Stars
This project builds a production-ready Click-Through Rate (CTR) prediction pipeline using real-world advertising data.
Convert Qwen3-Embedding-0.6B to ONNX format for Text Embeddings Inference (TEI)
Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 500+ LLMs (Qwen3, Qwen3-MoE, Llama4, GLM4.5, InternLM3, DeepSeek-R1, ...) and 200+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, Llava, GLM4v, Ph…
Zero-human, cold-start construction of long-chain agents in professional domains
Train speculative decoding models effortlessly and port them smoothly to SGLang serving.
Open-source implementation of AlphaEvolve
A flexible, adaptive classification system for dynamic text classification
Internal Coherence Maximization (ICM): A Label-Free, Unsupervised Training Framework for LLMs
DSPy: The framework for programming—not prompting—language models
Latest AI Jailbreak Payloads & Exploit Techniques for GPT, QWEN, and all LLM Models
DeepTeam is a framework to red team LLMs and LLM systems.
AmpleGCG: Learning a Universal and Transferable Generator of Adversarial Attacks on Both Open and Closed LLM
Official implementation of paper: DrAttack: Prompt Decomposition and Reconstruction Makes Powerful LLM Jailbreakers
[ArXiv 2025] Imperceptible Jailbreaking against Large Language Models
A productionized greedy coordinate gradient (GCG) attack tool for large language models (LLMs)
HexStrike AI MCP Agents is an advanced MCP server that lets AI agents (Claude, GPT, Copilot, etc.) autonomously run 150+ cybersecurity tools for automated pentesting, vulnerability discovery, bug b…
Two text jailbreak attacks against commercial black-box LLMs and a malicious content detection method, the latter of which is applied to red team dataset cleaning and jailbreak response detection.
Open-source implementation of an agent-centric red-teaming scanner inspired by Cloud Security Alliance guidance. Automated security testing for AI agents and LLM applications.
A sophisticated red-teaming agent built with LangGraph and Ollama to probe OpenAI's GPT-OSS-20B model for vulnerabilities and harmful behaviors. (Specifically built for the OpenAI Open Model Hackat…
A comprehensive dataset for Large Language Model (LLM) security evaluation, featuring three categories: Benign, Borderline, and Malicious. This repository provides critical data support for AI safe…
A.I.G (AI-Infra-Guard) is a comprehensive, intelligent, and easy-to-use AI Red Teaming platform developed by Tencent Zhuque Lab.
Rewrite to Jailbreak: Discover Learnable and Transferable Implicit Harmfulness Instruction (ACL2025)
The official implementation of our ICLR 2025 paper "One Model Transfer to All: On Robust Jailbreak Prompts Generation against LLMs".