Thanks for your interest in STYX. Here's how you can contribute.
This repository contains benchmark results and methodology for STYX (Selective Token Yield eXtraction). It does not contain the STYX algorithm itself, which is patent-pending (#63/975,190).
Found a flaw in the benchmark design? Have suggestions for stronger validation? Open an issue with the methodology label. We take statistical rigor seriously.
The benchmark files describe what was tested and how. If you run similar comparisons using the same public data sources (GitHub API, HuggingFace datasets) and get different results, we want to know. Open an issue with your findings.
Want to see STYX tested on a specific document type, language, or scale? Open an issue with the benchmark-request label.
If you spot an error in the benchmark JSON files or README, open an issue or submit a PR.
- Algorithm contributions — The core extraction algorithm is proprietary and patent-pending
- Requests for source code — The algorithm is not in this repo and will not be shared
- Raw + compressed text comparisons — We do not publish input/output comparisons to protect IP
Open an issue with the question label, or reach out at contact@lightspeedup.com.
Be respectful. Technical disagreement is welcome. Personal attacks are not.