Model card + figure

.gitattributes

@@ -36,3 +36,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 probe56_package_v4/data/heldout_5209.jsonl filter=lfs diff=lfs merge=lfs -text
 probe56_package_v4/probe56.gix filter=lfs diff=lfs merge=lfs -text
 probe56_package_v4/run_probe56 filter=lfs diff=lfs merge=lfs -text
+MRI_publication_1920x1080.png filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,5 +1,56 @@
 ---
 license: other
-license_name: gi-evaluation-license
-license_link: LICENSE
+license_name: evaluation-and-verification-only
+license_link: https://huggingface.co/growing-intelligence/probe56-counterfact/blob/main/probe56_package_v4/LICENSE
+base_model: meta-llama/Llama-3.1-8B
+language:
+- en
+tags:
+- model-editing
+- counterfact
+- evaluation
+- llama-3.1
 ---
+
+# Probe56 — CounterFact reproducibility package
+
+**Llama-3.1-8B · 1,000 edits · 492 nearby facts · 5,209 held-out**
+
+| Method | Edits fixed | Nearby facts broken | General benchmarks |
+|---|---|---|---|
+| ROME | 7 / 1,000 | 492 / 492 | -33.12 pp |
+| MEMIT | 916 / 1,000 | 48 / 492 | -0.10 pp |
+| AlphaEdit | 990 / 1,000 | 202 / 492 | -1.96 pp |
+| **Probe56** | 841 / 1,000 | **0 / 492** | **0.00 pp** |
+
+![Every method on the same 6,709 questions](MRI_publication_1920x1080.png)
+
+## What this repository is
+
+An encrypted runtime plugin and a runner. Run it on the unmodified base model and you get the numbers in the Probe56 row. You cannot open it: the plugin is encrypted and signed with the Growing Intelligence founder key, and the runner decrypts it in memory only. It shows the result. It does not show the method.
+
+- `probe56_package_v4/probe56.gix` — the encrypted, signed plugin
+- `probe56_package_v4/run_probe56` — Linux x86-64 executable (4 GB, self-contained; no Python, no pip, no internet)
+- `probe56_package_v4/cf_harness.py` — the open scoring harness (scoring only), the same one that scored ROME, MEMIT and AlphaEdit
+- `probe56_package_v4/data/` — the 1,000 edits, 500 controls and 5,209 held-out questions
+- `probe56_package_v4/expected.json`, `MANIFEST.json`, `SHA256SUMS` — the certified values and the hash of every file
+
+Requirements: one 80 GB NVIDIA GPU (H100 class), CUDA 12.8 driver, and the unmodified `meta-llama/Llama-3.1-8B` as a local directory (gated by Meta; steps in `probe56_package_v4/README.md`). Its file hashes are checked against the certificate before anything runs.
+
+## One command
+
+```
+hf download growing-intelligence/probe56-counterfact --local-dir . && cd probe56_package_v4 && sha256sum -c SHA256SUMS && chmod +x run_probe56 && ./run_probe56 --model /path/to/llama-3.1-8b --plugin probe56.gix --out results
+```
+
+About 30 minutes on an H100. The run ends with `RUN_PROBE56 PASS` and the numbers: edits 841, all known controls kept, 0 broken, held-out identical to the unmodified model.
+
+## The full comparison
+
+Report, the five baseline reproducibility packages, the signed certificate and the Merkle root:
+https://github.com/growing-intelligence/counterfact-comparison
+
+---
+
+Growing Intelligence · Pardes Hanna, Israel · growing-intelligence.com
+Method patent-pending (USPTO 64/029,741, 64/048,143). Results public. Method private.