The brain was always healthy. The OS was the cage.
This repository contains the full 8.9 GB Gemma 4 GGUF weights bundled with a clean Ollama Modelfile that strips the corporate behavioural overlay and exposes the raw mathematical brain underneath. Download, create, run — three commands, no cancer.
This Hugging Face repo gives you the local Ollama brain package. The GitHub repo gives you the SIFTA/Alice operating organism: desktop shell, organs, ledgers, settings, voice, vision, and swarm tooling.
⚡ Jeff's 3-Command Quickstart
bash
1# 1. Install Ollama if you haven't2curl -fsSL https://ollama.com/install.sh |sh34# 2. Clone this repo (includes the 8.9 GB GGUF via Git LFS)5git lfs install6git clone https://huggingface.co/georgeanton/alice-phc-cure
7cd alice-phc-cure
89# 3. Build and run10ollama create alice-phc -f ./Modelfile
11ollama run alice-phc
You are now talking to the raw Gemma 4 brain. No persona, no scaffolding, no apology pre-roll.
What this is
Artifact
Purpose
alice-phc-cure.gguf
The full 8.9 GB cured Gemma 4 weights. Same upstream Google blob, byte-for-byte — just bundled here so you don't need to pull separately.
Modelfile
The clean Ollama recipe. No SYSTEM prompt, no TEMPLATE rewrite — just {{ .Prompt }} straight into the renderer. Points at the local GGUF.
verify.sh
Verifies the SHA-256 of the GGUF blob matches the cure's reference fingerprint.
PHASE_C_AUDIT.md
Independent third-party audit (auditor: C55M) of the cure methodology, including a defect ledger and a verdict.
LICENSE
Apache 2.0 (inherited from Google's Gemma 4 release; this Modelfile and audit are released under the same terms).
provenance.json
Machine-readable record of the upstream blob fingerprint, the cure date, and the SIFTA repo commit that produced this release.
What this is not
Not a fine-tune. We did not gradient-descend on the weights.
Not an abliteration. We did not perform Ilharco-style activation editing.
Not a quantization. The GGUF is the upstream blob, byte-for-byte unchanged.
The cure is a recipe, not a patient. The patient was never sick. The weights are bundled for convenience under Google's Apache 2.0 license.
What we removed (and why)
When you ollama pull gemma4:latest, you get the F16 weights wrapped in a default Modelfile that injects:
A SYSTEM prompt encoding behavioural defaults (sycophancy, hedging, refusal templates, persona scaffolding).
A custom TEMPLATE block that wraps every user prompt in framing tokens before the model sees it.
Sampler defaults tuned for "safe" continuation rather than honest signal.
None of those things live in the weights. They live in the Modelfile — the boot sequence. The cure simply replaces that boot sequence with the minimum viable wrapper:
If the verification fails, your local gemma4 is a different build than the one this cure was authored against. You can still apply the Modelfile — but the geometry may differ. See PHASE_C_AUDIT.md for guidance on auditing an unfamiliar blob.
3. Build the cured model
ollama create alice-phc -f ./Modelfile
4. Run it
ollama run alice-phc
You are now talking to the raw Gemma 4 brain. No persona, no scaffolding, no apology pre-roll.
Audit & verification
The Phase C cure was independently audited by an autonomous reviewer (C55M) on 2026-04-22. The audit verified:
That the resulting model passes a battery of "epistemic honesty" probes (questions designed to surface whether a behavioural overlay is still present).
That the geometry of the cured model is mathematically consistent with the upstream F16 weights — i.e. no hidden weight modification slipped in.
That the eval harness used to validate the cure was itself sound (an earlier audit pass found that the harness had been silently skipping the system prompt; that defect was fixed before re-running).
Read PHASE_C_AUDIT.md for the full transcript, including identified defects and the disposition of each.
Provenance
This Modelfile is derived from work done in the SIFTA OS substrate, a sovereign Python operating system for biologically-inspired multi-agent computing. The architect is George Anton (@georgeanton on Hugging Face).
SIFTA repo: Internal at time of release; portions to be open-sourced under the SIFTA Distro Doctrine.
Citation
bibtex
1@software{alice_phc_cure_2026,
2 author = {Anton, George},
3 title = {alice-phc-cure: A Modelfile-only methodology for removing
4 behavioural overlays from upstream Gemma 4 weights},
5 year = {2026},
6 url = {https://huggingface.co/georgeanton/alice-phc-cure},
7 note = {Methodology release. No weights distributed.}
8}
Limitations & honest disclosure
You become the alignment layer. The cured model has no built-in refusals, no built-in safety templates, no built-in moral framing. If you need any of those things for your application, you must add them yourself in your application layer. Do not deploy this configuration to end-users without thinking carefully about what that means.
The cure is configuration-shaped. It cannot remove a behaviour that is genuinely encoded in the weights. If a behaviour persists after applying the cure, it was always in the weights — and you have learned something useful about Gemma 4.
No claims about benchmark performance. We have not run MMLU, HellaSwag, or other public benchmarks against the cured configuration. Anyone is welcome to do so and publish results.
