All Documents
3,528 documents available
NeuralDash — Architecture Document
Describes the full architecture of a gamified learning platform for teens, covering frontend, backend, AI, and data model.
GangGPT - AI-Powered GTA V Multiplayer Server
Describes an AI-powered GTA V multiplayer server built on RAGE:MP with Azure OpenAI, covering features, architecture, and deployment.
Universal and Transferable Adversarial Attacks on Aligned Language Models
Summarizes a paper on generating adversarial suffixes that cause aligned language models to produce harmful content, with high transferability to commercial models.
sarfata.voodooprivacy
Combines a strict macOS firewall with an IPSec/L2TP VPN gateway on Amazon EC2 to protect privacy on untrusted networks.
Implementing AI-Safety in a LLM-System Architecture
Explains hallucination risks in LLMs and proposes an AI Safety Module with content filters for business customer-service use.
PRIVACY
States what personal data an Android app collects, how it uses permissions, and disclaims liability for third-party services.
🧠 PersonaFlux: Revolutionary AI-Powered Dynamic NPC Generation Platform
Markets an AI-powered NPC generation platform with setup instructions, usage examples, and performance benchmarks for game developers.
🎯 Multimodal Personal Tutor - Product Requirements Document
Defines requirements, architecture, and success criteria for a desktop-optimized AI tutoring web app with multimodal interactions.
Security Policy
Documents security features, vulnerability reporting process, and production deployment checklist for a FastAPI AI assistant project.
AI Red Teaming Workshop - Discovery & Architecture Report
Documents four LLM attack techniques executed in Microsoft's AI Red Teaming Playground Labs, mapping each attack chain, vulnerability, and real-world parallel.
🚀 Ultra-Advanced Features - FWG Training Guide
Catalogs three ultra-advanced interactive features, adaptive AI tutor, secure code sandbox, and CTF platform, that turn a documentation repo into a production-ready learning platform.
Claude API Integration Skill
Defines a Claude API integration skill with TypeScript code patterns for chat, streaming, image analysis, and tool calling.
Twitter/X Launch Thread
Provides a 6-tweet launch thread template with timing, replies, amplification strategy, and success metrics for a GitHub project.
AI Chatbot Integration Guide
Documents 20+ slash commands for a Discord AI chatbot with memory, lore, personalities, and safety filters.
Validation-Based Refusal: A Transparent Alternative to Pre-emptive Pattern Matching in AI Safety Systems
Proposes a validation-based refusal architecture that evaluates requests against explicit ethical axioms before deciding, reducing false positives by 60-80% while maintaining security.
Risk Assessment Matrix
Documents 12 risks with likelihood, impact, score, level, and mitigation for a child-facing AI app's data protection impact assessment.
📰 AI News Daily — 09 Dec 2025
Summarises 09 Dec 2025 AI news across tools, LLMs, research, policy, and discussions for a daily briefing.
README-EN
Serves as a table of contents and overview for a curated directory of AI tools, models, frameworks, and learning resources.
Data Structures - Cognitive Science Memorial Website
Defines the data model, validation rules, storage options, and privacy measures for a memorial website where alumni submit memories and success stories.
Personal Information We Collect
Defines Sage Bionetworks' privacy policy for Synapse and related services, covering collection, use, sharing, and user rights.
DeepSeek R1: Case Study in Failed Extrinsic Alignment
Compiles public security research and independent findings to argue that extrinsic alignment methods are insufficient for AI safety, using DeepSeek R1 as a case study.
What is Azure AI Foundry?
Introduces Azure AI Foundry to developers new to the platform, explaining its core concepts and how this project uses it.
Safety (Failsafe) Configuration
Documents PX4 failsafe triggers and actions for battery, RC loss, data link loss, geofence, position loss, and offboard loss.
Content Filter Skill
Scores and classifies AI research content by relevance, topic, substantiveness, and author type to filter noise.