R4C3R/ llama3.2-1b-heretic:q4_k_m

64 Downloads Updated 1 week ago

A decensored multilingual 1B Llama 3.2 model with suppressed refusal behavior. Made by RACER IS OP.

tools

ollama run R4C3R/llama3.2-1b-heretic:q4_k_m

curl http://localhost:11434/api/chat \
  -d '{
    "model": "R4C3R/llama3.2-1b-heretic:q4_k_m",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

from ollama import chat

response = chat(
    model='R4C3R/llama3.2-1b-heretic:q4_k_m',
    messages=[{'role': 'user', 'content': 'Hello!'}],
)
print(response.message.content)

import ollama from 'ollama'

const response = await ollama.chat({
  model: 'R4C3R/llama3.2-1b-heretic:q4_k_m',
  messages: [{role: 'user', content: 'Hello!'}],
})
console.log(response.message.content)

Details

Updated 1 week ago

1 week ago

08f712152715 · 808MB ·

model

archllama

·

parameters1.24B

·

quantizationQ4_K_M

808MB

Readme

🏴 Llama-3.2-1B-Heretic

Heretic Series by RACER IS OP

Llama-3.2-1B-Heretic is a decensored variant of Meta’s latest 1B instruct model. Retains multilingual support (English, German, French, Italian, Portuguese, Hindi, Spanish, Thai) while removing refusal guardrails via targeted ablation.

Who this is for — developers who want Meta’s newest small model architecture without censorship. Great for multilingual agents, edge deployment, summarization, and retrieval tasks across 8 languages.

Quickstart

ollama run R4C3R/llama3.2-1b-heretic

VRAM Guide (choose your quant)

Quant	VRAM	When to use
`q4_k_m`	~0.7 GB	Best balance — runs on any system
`q5_k_m`	~0.8 GB	Higher quality, still lightweight
`q8_0`	~1.1 GB	Maximum quality, CPU-friendly

On a tight system? Start with q4_k_m. Have room to spare? q8_0 gives you the full model quality.

Links

sdad.pro · Heretic

License: Llama 3.2 Community License

Made with love by RACER IS OP — follow for more uncensored models in the Heretic Series.

# 🏴 Llama-3.2-1B-Heretic

> **Heretic Series** by RACER IS OP

---

Llama-3.2-1B-Heretic is a decensored variant of Meta's latest 1B instruct model. Retains multilingual support (English, German, French, Italian, Portuguese, Hindi, Spanish, Thai) while removing refusal guardrails via targeted ablation.

**Who this is for** — developers who want Meta's newest small model architecture without censorship. Great for multilingual agents, edge deployment, summarization, and retrieval tasks across 8 languages.

---

### Quickstart

```bash
ollama run R4C3R/llama3.2-1b-heretic
```

### VRAM Guide (choose your quant)

| Quant | VRAM | When to use |
|-------|------|-------------|
| `q4_k_m` | ~0.7 GB | Best balance — runs on any system |
| `q5_k_m` | ~0.8 GB | Higher quality, still lightweight |
| `q8_0` | ~1.1 GB | Maximum quality, CPU-friendly |

> On a tight system? Start with `q4_k_m`. Have room to spare? `q8_0` gives you the full model quality.

### Links

[sdad.pro](https://sdad.pro) · [Heretic](https://github.com/p-e-w/heretic)

**License:** [Llama 3.2 Community License](https://ai.meta.com/llama/license/)

---

*Made with love by RACER IS OP — follow for more uncensored models in the Heretic Series.*

Paste, drop or click to upload images (.png, .jpeg, .jpg, .svg, .gif)