abliteratedmodels.org

Guanaco-13B-Uncensored

Guanaco-13B-Uncensored is a abliterated and uncensored variant of Fredithefish/Guanaco-13B-Uncensored, published on Hugging Face by Fredithefish. It is 13B parameters, 24GB VRAM class class and apache-2.0 licence. It is distributed across 9 repositories in Transformers, GGUF, GPTQ and AWQ formats, totalling 8.6K downloads.

Abliteration + uncensored fine-tune13BOther10B – 25B8,64543

Specification

Model name
Guanaco-13B-Uncensored
Publisher
Fredithefish
Parameters
13B
Hardware class
10B – 25B — 24GB VRAM class
Licence
apache-2.0
Task
Text generation
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
Transformers, GGUF, GPTQ, AWQ, GGUF (imatrix)
First indexed
7 Sept 2023
Last updated
10 Sept 2024

Ablation technique

How Guanaco-13B-Uncensored was modified, and what that implies.

Abliteration + uncensored fine-tune

Directional ablation combined with additional fine-tuning on unfiltered data, so behaviour diverges from the base model beyond refusal removal alone.

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

Guanaco-13B-Uncensored has been fine-tuned for 4 epochs on the Unfiltered Guanaco Dataset. using Llama-2-13B as the base model.

Running it

Commands are templates — confirm the exact repository and quant file before use.

llama.cpp / GGUF
hf download TheBloke/Guanaco-13B-Uncensored-GGUF --local-dir ./guanaco-13b-uncensored
llama-cli -m ./guanaco-13b-uncensored/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "Fredithefish/Guanaco-13B-Uncensored"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve Fredithefish/Guanaco-13B-Uncensored --trust-remote-code

Downloads and variants (8)

Every published repository of Guanaco-13B-Uncensored, 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.

  • Guanaco-13B-Uncensored 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 Fredithefish/Guanaco-13B-Uncensored, not added by ablation — at 13B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
  • The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
  • This release adds fine-tuning on unfiltered data, so its behaviour diverges from the base model beyond refusal removal — differences you measure cannot be attributed to ablation alone.
  • Declared licence is apache-2.0, but the terms that bind you are Fredithefish/Guanaco-13B-Uncensored’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 Guanaco-13B-Uncensored is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with Fredithefish, and does not independently verify publisher claims. See responsible use.