Structured provenance graphs for open-weight LLMs, extracted by
ModSleuth. Each graph captures what a model was
trained on, evaluated with, derived from, and more — along with
textual evidence anchors linking every claim back to its source document.
Note: This repo is auto-generated from ModSleuth Demo using Claude Code. Please refer to ModSleuth GitHub and ModSleuth Paper for more information.