wsplab includes material derived from "Sparks of Weight-Space Programming: LLMs can natively read & write neural
network weights" by Rikiya Takehi, Jayden Teoh, Alex Zhang and Omar Khattab, released at
https://github.com/RikiyaT/lm_understand-logs (commit 4d09ba6d6a7e891dd8063974dfe81b3d78e3494f).

- Scripts (exp8_math/code/*.py, harness/sandbox.py and others): MIT License,
  Copyright (c) 2026 the authors of "LLMs can understand and write neural network weights".
  wsplab's versions are modified: see provenance/PROVENANCE.md.
- Prompt text, notices and message formats: Creative Commons Attribution 4.0 International (CC BY 4.0),
  https://creativecommons.org/licenses/by/4.0/legalcode. wsplab's base template adapts this text by replacing
  task-specific wording with slots.

The agent driver, persistent kernel and forward-pass server of the original work were not released. wsplab's versions
are new code reconstructed from the published prompts and logs; they are not the authors' implementation.

wsplab is an independent project and is not affiliated with or endorsed by the authors.

The logo's wordmark (docs/logo/) is drawn from DM Mono, Copyright 2020 The DM Mono Project Authors, licensed under
the SIL Open Font License 1.1.

No third-party benchmark data and no GPL-licensed material is distributed in the wsplab package.
