A standardised malware analysis report template with a JSON Schema and Markdown format mapped to MITRE ATT&CK. Validate, share and compare reports consistently. Part of ThreatBytes.
Accessible large language models via k-bit quantization for PyTorch.
Single-GPU LLM fine-tuning preflight: memory estimates, runnable LoRA/QLoRA recipes, and local measurements.
A small proof-of-concept language model (not an LLM) that aims to incorporate latent-space prediction, internal state using recurrent trace units, and byte-by-byte output, built with MLX.