ranger
51bc7d88cf
Matches the README pattern used across other top-level folders (modelfiles, observations, paper, etc.). Provides examiners with a quick entry point into the documentation: what each file is, recommended reading order, and why these docs exist alongside the primary research artefacts. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2.0 KiB
2.0 KiB
Documentation — Research Context & Project Journey
Curated documentation that supports the CyberRanger CA1 thesis. Includes the complete version evolution, mirrored research blog posts, and an index linking back to the live blog.
What's Here
| File / Folder | Description | Read Time |
|---|---|---|
version-evolution.md |
Complete V1 → V43 journey across 6 eras — Genesis, Exploration, Refinement, Production Hardening, Architecture Maturation, QLoRA Validation. 40+ versions documented with quick-reference table, per-version detail, and key-lessons-by-era summary. Used as Appendix C in the CA1 paper. | 15 min |
research-blog.md |
Curated index of 6 research blog posts with both live URLs and offline-mirror links. Citation format included for academic reuse. | 3 min |
blog/ |
Offline mirrors of 6 research blog posts (point-in-time snapshots). The live blog at https://davidtkeane.github.io is the canonical source; these copies are frozen for academic record. | — |
Recommended Reading Order
- Start here:
version-evolution.md— gives the full journey from dental chatbot (V1) to QLoRA-validated production model (V42-gold-wrapped) in 15 minutes. - Then:
research-blog.md— index page that points to the underlying experimental write-ups. - Deep dive:
blog/2026-03-08-from-rangerbot-to-cyberranger-v42-the-full-story.md— long-form narrative of the same journey.
Why These Docs Are Here
The Modelfiles, training data, and evaluation scripts are the primary research artefacts. The documents in this folder provide the interpretive layer — the journey, the failed iterations, the regressions, the breakthroughs — that transforms 86 Modelfiles into a coherent research story.
For examiners, this is the fastest path into the project without reading every Modelfile.