Machine Learning – core-jmp

// Machine Learning

3 posts

From P-Code to GNN: Extracting Binary Code Semantics

Quarkslab’s pcode_graph lifts functions to Ghidra P-Code CDGs and a GINE embedding that matches GMN (AUC 0.87 XM) on Cisco-Talos, across ISA,…

oxfemale· Sep 14· 15 min ·474

How LLMs Actually Work: A Transformer Internals Walkthrough

AI Security Research585

AI Security Research LLM

How LLMs Actually Work: A Transformer Internals Walkthrough

0xkato's walkthrough of the mechanisms inside modern transformer-based LLMs — tokenization, embeddings, Rotary Position Embeddings, attention with Q/K/V and causal masking, multi-head…

oxfemale· Jun 22· 29 min ·585

Abusing WinML for In-Memory Staging and EDR Evasion

Bypassing371

Bypassing cpp

Abusing WinML for In-Memory Staging and EDR Evasion

The research shows how attackers can embed payloads inside ONNX ML models and load them via Windows WinML APIs, staging malware entirely…

oxfemale· Apr 14· 23 min ·371