SZLHOLDINGS/szl-1-doctrine-sft
Part of the SZL Holdings governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance. SZL-1 Doctrine SFT The supervised fine-tuning (SFT) set that teaches SZL-1 its identity and SZL Holdings' honesty doctrine. This is the exact training data used by the szl-forge kit (train_szl.py, Unsloth QLoRA on unsloth/Qwen2.5-3B-Instruct-bnb-4bit). What's in it — honest labels… See the full description on the dataset page: https://huggingface.co/datasets/SZLHOLDINGS/szl-1-doctrine-sft.
<!-- SZL-ESTATE-CARD:v2:START --> <p align="center"><a href="https://a-11-oy.com/"><img src="https://huggingface.co/spaces/SZLHOLDINGS/README/resolve/main/assets/estate-banner-v2.svg" alt="SZL Holdings — governed, receipted, verifiable" width="100%"></a></p> <p align="center"> <a href="https://github.com/szl-holdings/.github/tree/main/doctrine"><img src="https://img.shields.io/badge/doctrine-v11%20LOCKED-0B1F3A?style=flat-square" alt="doctrine v11"></a> <a href="https://a-11-oy.com/"><img src="https://img.shields.io/badge/evidence%20wall-LIVE%20%C2%B7%20verify%20in%20browser-3AF4C8?style=flat-square" alt="live evidence wall"></a> <a href="https://huggingface.co/datasets/SZLHOLDINGS/szl-lake"><img src="https://img.shields.io/badge/szl--lake-offline%20verifiable-C9B787?style=flat-square" alt="szl-lake offline verifiable"></a> <a href="https://huggingface.co/spaces/SZLHOLDINGS/holographic"><img src="https://img.shields.io/badge/estate%20map-holographic-5B8DEE?style=flat-square" alt="holographic estate map"></a> </p> <p align="center"><sub>Part of the <a href="https://huggingface.co/SZLHOLDINGS">SZL Holdings</a> governed estate — claims are designed to carry checkable receipts. Verification proves integrity & origin, never accuracy or performance.</sub></p> <!-- SZL-ESTATE-CARD:v2:END -->
<div align="center"> <p>
  
</p> </div>
SZL-1 Doctrine SFT
The supervised fine-tuning (SFT) set that teaches SZL-1 its identity and SZL Holdings' honesty doctrine. This is the exact training data used by the `szl-forge` kit (train_szl.py, Unsloth QLoRA on unsloth/Qwen2.5-3B-Instruct-bnb-4bit).
What's in it — honest labels first
Evaluation
None yet. No eval, quality, speed, or accuracy numbers have been measured on SZL hardware. Treat any performance of a model trained on this set as UNKNOWN until a real run produces MEASURED numbers. No benchmark number is stated here because none exists — per SZL's honesty doctrine, none may be invented.
Reproduce
The training hyperparameters are pinned in `szl-forge/config.yaml` (seed 11, 3 epochs, LoRA rank 16, lr 2e-4). Run the one-command bootstrap from the szl-forge README, or python train_szl.py with this file as szl_dataset.jsonl in the working directory.
Built and maintained by SZL Holdings.
<div align="center">
[🛡️ SZLHOLDINGS on Hugging Face →](https://huggingface.co/SZLHOLDINGS) · [a-11-oy.com →](https://a-11-oy.com) · [Estate hub — live →](https://szlholdings-szl-estate-live.static.hf.space)
Governed AI you can prove.
<sub>SLSA: L1 honest · L2 attested · L3 roadmap. Λ = Conjecture 1 (advisory, never a theorem). Trust ceiling 0.97 — never 100%. Labels honest by default: MEASURED / REPORTED / MODELED / HEURISTIC / UNKNOWN / UNAVAILABLE. locked-proven = exactly 8 {F1,F4,F7,F11,F12,F18,F19,F22}.</sub>
</div>