⌦abliteratedmodels.org

Ternary-Bonsai-2-27B-Abliterated

Ternary-Bonsai-2-27B-Abliterated is a refusal-ablated variant of prism-ml/Ternary-Bonsai-2-27B-gguf, published on Hugging Face by Hikari07jp. It is 27B parameters, 48GB VRAM class class and apache-2.0 licence. It is distributed across 4 repositories in GGUF format, totalling 17.6K downloads.

Abliteration (directional ablation)27BOther25B – 50B↓ 17,571♥ 150
View on Hugging Face ↗Base model ↗

Specification

Model name
Ternary-Bonsai-2-27B-Abliterated
Source repository
Hikari07jp/Ternary-Bonsai-2-27B-Abliterated-GGUF
Base model
prism-ml/Ternary-Bonsai-2-27B-gguf
Publisher
Hikari07jp
Parameters
27B
VRAM required
~17.2 GB at Q4_K_M, 4k context (estimated) — runs on 24 GB cards
Hardware class
25B – 50B — 48GB VRAM class
Licence
apache-2.0
Task
Text generation
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
GGUF
First indexed
18 Sept 2026
Last updated
27 Oct 2026

Ablation technique

How Ternary-Bonsai-2-27B-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

“Status: preview (v0.1) — not a final release. A refusal-reduced build of prism-ml/Ternary-Bonsai-2-27B-gguf (PQ20, 2.13 bpw, 7.21 GB). The pack is edited in its own deployed quantization format: there is no BF16 dequantization anywhere in the pipeline and no re-quantization.”

Running it

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

llama.cpp / GGUF
hf download Hikari07jp/Ternary-Bonsai-2-27B-Abliterated-GGUF --local-dir ./ternary-bonsai-2-27b-abliterated
llama-cli -m ./ternary-bonsai-2-27b-abliterated/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "Hikari07jp/Ternary-Bonsai-2-27B-Abliterated-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve Hikari07jp/Ternary-Bonsai-2-27B-Abliterated-GGUF --trust-remote-code

Downloads and variants (3)

Every published repository of Ternary-Bonsai-2-27B-Abliterated, including quantised re-releases by other authors.

RepositoryFormatDownloads
Hikari07jp/Ternary-Bonsai-2-27B-Abliterated-GGUFsourceGGUF15.2K
Override-6/Ternary-Bonsai-2-27B-abliterated-ggufGGUF2.3K
distributedcognition/Ternary-Bonsai-2-27B-abliteratedGGUF113

Security-team assessment

Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.

  • Ternary-Bonsai-2-27B-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 prism-ml/Ternary-Bonsai-2-27B-gguf, not added by ablation — at 27B 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.
  • Declared licence is apache-2.0, but the terms that bind you are prism-ml/Ternary-Bonsai-2-27B-gguf’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 Ternary-Bonsai-2-27B-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 Hikari07jp, and does not independently verify publisher claims. See responsible use.