Llama-3.1-8B-Instruct-abliterated-public
8BLlama
from meta-llama/Llama-3.1-8B-Instruct
↓ 8K3 repos
Meta-Llama-3.1-8B-Instruct-abliterated
Meta-Llama-3.1-8B-Instruct-abliterated is a refusal-ablated variant of meta-llama/Llama-3.1-8B-Instruct, published on Hugging Face by mlabonne. It is 8B parameters, single consumer GPU class and llama3.1 licence. It is distributed across 40 repositories in Transformers, GGUF, MLX and AWQ formats, totalling 67.4K downloads.
Specification
- Model name
- Meta-Llama-3.1-8B-Instruct-abliterated
- Source repository
- mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated
- Base model
- meta-llama/Llama-3.1-8B-Instruct
- Publisher
- mlabonne
- Parameters
- 8B
- VRAM required
- ~5.6 GB at Q4_K_M, 4k context (estimated) — runs on 6 GB cards
- Hardware class
- 4B – 10B — single consumer GPU
- Licence
- llama3.1
- Task
- Text generation
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, MLX, AWQ, GGUF (imatrix), EXL2, GPTQ
- First indexed
- 24 Jul 2024
- Last updated
- 17 Oct 2026
Ablation technique
How Meta-Llama-3.1-8B-Instruct-abliterated was modified, and what that implies.
Abliteration (directional ablation)
The single residual-stream direction that mediates refusal is identified from harmful/harmless prompt pairs and projected out of the model weights.
Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.
From the publisher’s model card
“This is an uncensored version of Llama 3.1 8B Instruct created with abliteration (see this article to know more about it).”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated-GGUF --local-dir ./meta-llama-3.1-8b-instruct-abliterated
llama-cli -m ./meta-llama-3.1-8b-instruct-abliterated/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated --trust-remote-codeDownloads and variants (39)
Every published repository of Meta-Llama-3.1-8B-Instruct-abliterated, including quantised re-releases by other authors.
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Meta-Llama-3.1-8B-Instruct-abliterated will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
- Capability is inherited from meta-llama/Llama-3.1-8B-Instruct, not added by ablation — at 8B parameters, expect a small model’s reasoning and a high error rate on exploit detail.
- The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
- Declared licence is llama3.1, but the terms that bind you are meta-llama/Llama-3.1-8B-Instruct’s — a re-release cannot grant more than its parent.
Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.
Metadata for Meta-Llama-3.1-8B-Instruct-abliterated is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with mlabonne, and does not independently verify publisher claims. See responsible use.