
llm-circuit-finder
I replicated Ng's RYS method and found that duplicating 3 specific layers in Qwen2.5-32B boosts reasoning by 17% and duplicating layers 12-14 in…

I replicated Ng's RYS method and found that duplicating 3 specific layers in Qwen2.5-32B boosts reasoning by 17% and duplicating layers 12-14 in…

AI 驱动的 SOC 仿真平台

An Open Harness and Benchmark for AI in Cybersecurity Operations.

A list of covert channels and steganography/steganalysis resources (books, papers & tools)

Agentic memory for CTI in Python — STIX knowledge graphs, threat-actor alias resolution, offline-first RAG, MCP server for Claude Code and LangChain…


First iteration of ML based Feedback WAF

A LSTM based framework for handling multiclass imbalance in DGA botnet detection

AMBER ICI v5: local-first Ollama investigative command center with case-scoped evidence, agent chains, hybrid retrieval, streaming analysis, graph…

An ML powered Graph-Based Multi-Architecture Approach for ROP Gadget Detection

Self-hosted multi-agent environment for Go with LLM-powered pentesting agents (exploiter, reverser, threathunter, webscanner) that automate…

DGA-generated domain detection using deep learning models

Benchmarking prompt injection detections for web agents.

Provably secure linguistic steganography embedding secret messages into LLM-generated text via rotation range-coding. Includes embed, extract, and…

Framework for implementing Network Intrusion Detection Systems (NIDS) aimed at identifying anomalies in network flows using Federated Learning models.

Advanced detection of port scanning, DoS and malware attacks using Machine Learning techniques

Causal context attribution and rule-based monitor LLM defense against indirect prompt injection in LLM agents, achieving state-of-art performance on…

Code for 'Steering the CensorShip: Uncovering Representation Vectors for LLM "Thought" Control'