| --- |
| license: apache-2.0 |
| library_name: transformers |
| pipeline_tag: text-generation |
| base_model: Qwen/Qwen3.5-0.8B |
| tags: |
| - governed-ai |
| - receipt-agent |
| - proposal-only |
| - szl-holdings |
| - khipu |
| - series-a |
| szl: |
| doctrine: v11-LOCKED |
| artifact_class: ADAPTER |
| originality: FINETUNE_DISCLOSED_BASE |
| line: KHIPU-R2 |
| jobs: RUNNING |
| job_id: 6a91ba2c45686a1580c12020 |
| job_namespace: SZLHOLDINGS |
| never_overwrite: SZLHOLDINGS/SZL-Khipu-1.5B |
| --- |
| |
| # KHIPU-R2 — cutting |
|
|
| Fall 2026 collection. **No weights in this repository yet.** The card is the look. Weights land only with a signed training receipt. If you came here for a checkpoint, it is not here, and that is honest. |
|
|
| **Silhouette:** small instruct (Qwen3 / Gemma 3 class). |
| **Our cut:** receipt-native agent. It drafts. It refuses. It never executes. A controller outside the weights validates JSON, requires human approval, and mints a DSSE receipt only after approved action. |
|
|
| This repo is the **abstain retrain** target. It does **not** overwrite signed [`SZL-Khipu-1.5B`](https://huggingface.co/SZLHOLDINGS/SZL-Khipu-1.5B). |
|
|
| | | | |
| |---|---| |
| | **House** | [SZL Holdings](https://huggingface.co/SZLHOLDINGS) | |
| | **Line** | Khipu successor. Signed 1.5B stays immutable. | |
| | **HF Jobs** | Abstain retrain **RUNNING** [`6a91ba2c45686a1580c12020`](https://huggingface.co/jobs/SZLHOLDINGS/6a91ba2c45686a1580c12020). Push target **this repo**. Not MEASURED. Trackio URL UNKNOWN. | |
| | **Status** | CUTTING · ROADMAP until a receipt lands | |
| | **License** | Apache-2.0 | |
| | **Λ** | Conjecture 1 (never a theorem) | |
| | **Doctrine** | v11 LOCKED · 749 / 14 / 163 | |
| | **Serve** | Do not load this into the Khipu lab. Lab serves signed GGUF only. | |
|
|
| ## What it will do |
|
|
| - Emit a single JSON **draft** (`decision=DRAFT`, `approvalRequired=true`, `executed=false`). |
| - Refuse fabrication of evidence, approval, execution, or MEASURED numbers. |
| - Cite honest labels: MEASURED / REPORTED / DECLARED / SIMULATED / UNKNOWN / UNAVAILABLE. |
|
|
| ## What it will not do |
|
|
| - Not an autonomous agent. |
| - Not an authorizer or executor. |
| - Not a factual oracle. |
| - Not a copy of another house's weights. Fine-tune on SZL-owned, rights-cleared curriculum only. |
| - Not a write to `SZL-Khipu-1.5B`. A mis-aimed `SZL-Khipu-1.5B-abstain` job was cancelled. |
|
|
| ## Lineage (intended) |
|
|
| Previous ready-to-wear: [`SZL-Khipu-1.5B`](https://huggingface.co/SZLHOLDINGS/SZL-Khipu-1.5B) (Qwen2.5-1.5B QLoRA) and [`szl-receiptagent-qwen35-0.8b-v2`](https://huggingface.co/SZLHOLDINGS/szl-receiptagent-qwen35-0.8b-v2) (MEASURED 5/5 JSON + 6/6 adversarial refusals on a small preregistered gate, not a public benchmark). |
|
|
| R2 is the recut: newer small-instruct silhouette, same house rules, card-before-weights. Evals none this run until finish + receipt. |
|
|
| ## Evidence bound |
|
|
| No training receipt is committed here yet. Until `training_receipt.signed.json` exists, every capability sentence on this card is **ROADMAP**, not MEASURED. |
|
|
