"How to Tell If an LLM Was Really Trained From Scratch: A Reproducible Fingerprinting Method"

Model DNA, Analyzed: Verifying 'From-Scratch' LLM Claims with Architecture, Tokenizer, and CKA (PyTorch)

A technical deep-dive that cites and reproduces the Model DNA method in PyTorch - verifying 'trained from scratch' LLM claims via architecture config, tokenizer overlap, and…