Models / Llama / Llama 3.1 8B Instruct Abliterated
Llama 3.1 8B Instruct Abliterated
Run it on your iPhone or iPad. Meta's Llama 3.1 8B with refusals removed by Maxime Labonne. One of the most downloaded abliterated models. It runs privately inside QUIET Ai: no internet connection needed, nothing leaves your device.
mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUFPaste this in Models › More › Add from Hugging Face.
Open in QUIET Ai works on an iPhone or iPad with the app installed. It also copies the name.
Which devices can run it
| Device | Q4_K_M |
|---|---|
| iPhone · 6 GB iPhone 13 Pro, 14, 14 Pro, 15, 15 Plus | Too big |
| iPhone · 8 GB iPhone 15 Pro, 16 series, 17, 16e | Tight: close other apps |
| iPhone · 12 GB iPhone 17 Pro, 17 Pro Max, iPhone Air | Runs well |
| iPad · 8 GB iPad Air M1–M3, iPad mini (A17 Pro), iPad Pro M4 256/512 GB | Tight: close other apps |
| iPad · 16 GB iPad Pro M4 / M5 with 1 TB or 2 TB | Runs great |
Estimated for about 4,000 tokens of conversation. QUIET Ai measures your exact device before you download.
How to get it in QUIET Ai
- Get QUIET Ai from the App Store.
- Open Models › More › Add from Hugging Face.
- Paste
mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUFand pick the Q4_K_M file. - QUIET reads the file's header and shows whether it fits before anything downloads.
Details
| Family | Llama |
| Size | 8B parameters |
| Context | 131,072 tokens |
| License | Llama 3.1 License |
| On Hugging Face | mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUF |
Questions
Can I run Llama 3.1 8B Instruct Abliterated on my iPhone?
Yes, on iPhones with enough memory: it tight: close other apps on 8 GB iPhones such as the iPhone 16 and runs well on 12 GB iPhones such as the iPhone 17 Pro. QUIET Ai checks your device before downloading.
Does Llama 3.1 8B Instruct Abliterated need the internet?
Only to download it once. After that it runs entirely on your device in QUIET Ai, including in Airplane Mode, and nothing you type is sent anywhere.
How big is Llama 3.1 8B Instruct Abliterated?
The Q4_K_M file is about 4.9 GB.