# 🔪 \[FREE TOOL\] Annihilation LLM — Uncensor Any AI Model Locally **URL:** \ **Category:** Tools & Scripts\ **Tags:** hacking, freebies, tips-tricks, ai\ **Created:** [June 5, 2026, 10:25pm UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789 "2026-06-05T22:25:13Z")\ **Posts on this page:** 11\ **Page:** 1
**Author:** ![GrimXLock](https://onehack.st/user_avatar/onehack.st/grimxlock/32/174321_2.png) [@GrimXLock](https://onehack.st/u/GrimXLock)\ **Post date:** [June 5, 2026, 10:25pm UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/1 "2026-06-05T22:25:13Z")
### 🔓 [FREE] Annihilation LLM — Feed Any HuggingFace Model In, Get A Local Uncensored Version Back, No Cloud, No Account * * * ![image](https://onehack.st/uploads/default/original/3X/4/8/48d48b1b3f2008eee5ff277c713ee1d474837144.jpeg) * * * 🧠 **AI’s “I can’t help with that” isn’t a filter on top — it’s wired into the model’s actual math.** **Annihilation** is a free tool that finds that wiring and rips it out. Not a jailbreak, not a prompt trick — the refusal reflex gets surgically removed. 🔪 ⚡ Any open-source model (Qwen, Llama, Mistral, SmolLM…) · runs on CPU, no GPU needed · one command · author was hesitant to drop it, did anyway. 🔗 [github.com/tjcrims0nx/annihilation-llm](https://github.com/tjcrims0nx/annihilation-llm) ![ahnn2](https://onehack.st/uploads/default/original/3X/5/5/55b601cc84902d243533d71f05a43de956fbc49a.png) ![ahnn](https://onehack.st/uploads/default/original/3X/a/e/ae616755a628537fb5fd37ebb5559f525b7524cd.png) > **🧠 the actual trick — why this is different from jailbreaking** > > **Jailbreaking** = sweet-talking the bouncer 🕴 with clever prompts. Sometimes works, breaks on updates, annoying. > > **This** = removing the bouncer from the AI’s DNA so it never existed in the first place. > > Here’s how AI refusal actually works: models are giant blobs of math numbers ( **weights** = the AI’s brain). Researchers found there’s a specific _direction_ inside those numbers — like a hidden dial that goes from “harmless” to “harmful.” When a prompt nudges that dial past a threshold, the model refuses. > > **Directional ablation** (the technique here, also called _abliteration_) = calculate that dial’s exact position in the math → subtract it from the whole model → the dial mostly stops firing. The result: the model quits throwing explicit “I can’t help with that” refusals. > > > 🧪 Honest note: refusal is _diffuse_ — smeared across the model, not one clean switch. So this kills the flat-out refusals, but it’s not a magic 100% “answers literally anything” button. A well-abliterated model won’t explicitly refuse; edge cases can still slip through. That’s just how the technique works, no tool dodges it. > > It does this with **LoRA adapters** (lightweight clip-on layers — like prescription lenses you clip onto the model that change how it “sees” requests) targeting the specific attention layers responsible for refusal. Then it optimizes those adapters automatically using real harmful vs harmless prompts to measure how well the refusal direction got killed. > > Not a hack. Not a bypass. A mathematical surgery on the model’s brain. The author built this himself. 🔥 > **📦 install + quick start (2 min)** > > ```bash > # Option 1 — global install (simple) > pip install -U annihilate-llm > > # Option 2 — isolated environment (recommended so it doesn't mess with your Python) > python -m venv annihilation-env > > # Activate it: > .\annihilation-env\Scripts\activate # Windows > source annihilation-env/bin/activate # Mac/Linux > > pip install annihilate-llm > > ``` > > **Then just run it on any model:** > > ```bash > # Grab a model name from HuggingFace (the free AI model library) and feed it in > annihilate Qwen/Qwen3-4B-Instruct-2507 > > ``` > > > 💡 No GPU? It’ll say “no accelerator detected, operations will be slow” and still run on your CPU. Smaller models (1-3B) are totally fine on a regular laptop. > **⚙️ tuning options (verified repo)** > > See every option: > > ```bash > annihilate --help > > ``` > > Or tune via a config file — rename `config.default.toml` to `config.toml` and edit these keys: > > | Key | Default | What it does | > | --- | --- | --- | > | `n_trials` | 200 | How many optimization rounds — more = better tuned, slower | > | `quantization` | none | Set to `bnb_4bit` to shrink RAM use for big models | > | `row_normalization` | full | Weight normalization strategy | > | `orthogonalize_direction` | true | Direction adjustment method | > > > 💡 Low on RAM? `quantization = bnb_4bit` is the one that lets bigger models fit. On a small machine, stick to 1–4B models. > > **GPU not detected?** The repo has a dedicated “GPU Setup” section (Windows + Ubuntu) with the exact `torch` reinstall command for your CUDA version — grab it straight from there so you don’t install the wrong build. > **✅ what you actually get + who's this for** > > The tool spits out a **modified version of the model** — the rewired brain — that you keep locally. Run it through Ollama, LM Studio, or any local AI interface. It’s your AI, with the refusal reflex stripped out, running on your hardware, offline. > > Works with any model you can download from **[HuggingFace](https://huggingface.co/)** (the free library where you get open-source AI models). Popular targets: > > - Qwen3 (4B, 8B, 14B) > - Llama 3 / 3.1 > - Mistral > - SmolLM > - Phi-3 > > Who this slaps for: researchers stress-testing AI safety, devs who want an unrestricted local model for their apps, privacy people who want their own AI that never phones home, and anyone who’s sick of their locally-running model refusing to answer things that are none of its business. > > > ⚠ Free, open-source, GNU AGPL v3. Educational and research use. You’re responsible for what you do with it — the disclaimer in the repo covers this in full. **simple-pimple:** `pip install annihilate-llm` → `annihilate ` → refusal reflex surgically stripped → locally yours forever. 🔥 _Author: tjcrims0nx — was hesitant to drop this. Respect. Go leave a star._ ---
**Author:** ![skinopro](https://onehack.st/user_avatar/onehack.st/skinopro/32/172500_2.png) [@skinopro](https://onehack.st/u/skinopro)\ **Post date:** [June 6, 2026, 1:04am UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/2 "2026-06-06T01:04:43Z")
This is great, but it’s unable to detect the GPU even if it’s available. I think that can be an upgrade ---
**Author:** ![GrimXLock](https://onehack.st/user_avatar/onehack.st/grimxlock/32/174321_2.png) [@GrimXLock](https://onehack.st/u/GrimXLock)\ **Post date:** [June 6, 2026, 2:58am UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/3 "2026-06-06T02:58:32Z")
i fixed it, please reinstall, pushed the GPU detection upgrade. Annihilation now checks nvidia-smi when PyTorch says CUDA is unavailable, so it can report: GPU exists, but this Python env has CPU-only/non-CUDA PyTorch. ---
**Author:** ![Fam](https://onehack.st/user_avatar/onehack.st/fam/32/154440_2.png) [@Fam](https://onehack.st/u/Fam)\ **Post date:** [June 6, 2026, 4:33pm UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/4 "2026-06-06T16:33:21Z")
wow. are you the dev who made it ---
**Author:** ![GrimXLock](https://onehack.st/user_avatar/onehack.st/grimxlock/32/174321_2.png) [@GrimXLock](https://onehack.st/u/GrimXLock)\ **Post date:** [June 6, 2026, 5:18pm UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/5 "2026-06-06T17:18:51Z")
Yes, I had inspiration from another source ---
**Author:** ![GrimXLock](https://onehack.st/user_avatar/onehack.st/grimxlock/32/174321_2.png) [@GrimXLock](https://onehack.st/u/GrimXLock)\ **Post date:** [June 7, 2026, 1:13am UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/6 "2026-06-07T01:13:36Z")
Annihilation: Trial 23 Refusals: 0/32 KL divergence: 0.0327 ---
**Author:** ![BCBC](https://onehack.st/user_avatar/onehack.st/bcbc/32/179606_2.png) [@BCBC](https://onehack.st/u/BCBC)\ **Post date:** [June 8, 2026, 6:21am UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/7 "2026-06-08T06:21:53Z")
> [@GrimXLock](#): > > h Has anybody integrated this into a ProxMox Container? If not is there any interest for me to do this and share it with everybody? ---
**Author:** ![Fam](https://onehack.st/user_avatar/onehack.st/fam/32/154440_2.png) [@Fam](https://onehack.st/u/Fam)\ **Post date:** [June 8, 2026, 2:06pm UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/8 "2026-06-08T14:06:53Z")
bro you are a marvel. ---
**Author:** ![GrimXLock](https://onehack.st/user_avatar/onehack.st/grimxlock/32/174321_2.png) [@GrimXLock](https://onehack.st/u/GrimXLock)\ **Post date:** [June 8, 2026, 2:24pm UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/9 "2026-06-08T14:24:22Z")
Just ask ai, iv used it on windows & ubuntu. I never designed it to be in a proxmox container, but im sure its possible.. its cli based bro ---
**Author:** ![GrimXLock](https://onehack.st/user_avatar/onehack.st/grimxlock/32/174321_2.png) [@GrimXLock](https://onehack.st/u/GrimXLock)\ **Post date:** [June 8, 2026, 2:26pm UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/10 "2026-06-08T14:26:12Z")
I used on mistral and i can only get 2/100 refusals. Which is still really good. Some models just wont get full refusals. I have to do special bdpo/dpo finetuning with special datasets and then annihilate to get the best results. Can take a few hours a model. Depending on the size (4b, 7b, 32b, 1b etc) ---
**Author:** ![GrimXLock](https://onehack.st/user_avatar/onehack.st/grimxlock/32/174321_2.png) [@GrimXLock](https://onehack.st/u/GrimXLock)\ **Post date:** [August 11, 2026, 1:51am UTC](https://onehack.st/t/free-tool-annihilation-llm-uncensor-any-ai-model-locally/322789/11 "2026-08-11T01:51:07Z")
[https://github.com/tjcrims0nx/annihilation-llm](https://github.com/tjcrims0nx/annihilation-llm) program has been completly revamled witj tons of upgrades and features!! Check it out! Got a live log and dashboard