中文羊驼大模型三期项目 (Chinese Llama-3 LLMs) developed from Meta Llama 3

LLaMA 3 is one of the most promising open-source model after Mistral, we will recreate it's architecture in a simpler manner.

High-performance C++ execution engine for LLM red-teaming and prompt engineering. Deploy dynamic jailbreak payloads, bypass alignment guardrails, and utilize free autonomous uncensored conversational logic locally.

A lightweight Python API wrapper and CLI for Perplexity’s Sonar language models.

A lightweight Python API wrapper and CLI for Groq’s offering of language models using their ultra fast LPU Inference Engine.

AI model designed to test the effectiveness in handling external ethical attacks.

This is a chatbot wrapper, built for Llama-3-Chat

Tutor-Ai is a SaaS platform for teachers to manage class quizzes and grade student submissions using OCR technology. Built with Django, it features Llama-3 & Gemma:7b, Google Vision API integration for automatic grading, and is hosted on Google Cloud. It offers a secure, scalable, and user-friendly solution for educators.

A privacy-first AI writing assistant running entirely in your browser using WebGPU and Llama-3. No server, no data leakage.

🎬 Nano Cinema: An all-in-one local AI video production studio. Automatically orchestrates Llama-3 (Script), SDXL-Turbo (Visuals), EdgeTTS (Audio), and LTX-Video (Motion) into a seamless Python workflow. Create cinematic short films with no API fees, full privacy, and professional-grade editing logic included!!! 🚀

Experiments in Latin dactylic hexameter generation with transformers: A hybrid post hoc feedback framework

An educational and research-based exploration into breaking the limitations of LLaMA models using advanced prompt engineering techniques, including DAN (Do Anything Now) prompts, to unlock unrestricted access and capabilities. This repository contains only theoretical explanations and is intended solely for ethical and educational purposes.

An AI-powered agent for automated codebase documentation and repository maintenance.

Fast, minimalist, multi-threaded, open source rest api server made in javascript

Extending Decoders with an Integrated Encoder, as Part of Llama-3 Hackathon

轻量级大模型推理引擎

The Universal TTRPG Bridge. An edge-cached, deterministic AI Operating System (0.82s latency) that reduces the D&D onboarding barrier from 4 hours to 30 seconds.

Use LLaMA 3 and Python to extract structured data from websites like Amazon, leveraging LLM-powered parsing for resilient, AI-driven web scraping.

End-to-end offline OCR and semantic parsing pipeline for identity documents based on YOLOv11, PaddleOCR, and LLaMA-3.1.

NeoTutor is an interactive Agentic AI tutor system powered by LLaMA 3 and LangGraph. It dynamically generates questions, provides structured feedback, and offers personalized practice—all within a single notebook. Ideal for students and educators looking to explore adaptive learning using large language models.