⌦abliteratedmodels.org

Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO

Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO is a Heretic-ablated variant of ukisai/Swift-1.5-Qwen3.8-27b, published on Hugging Face by jegreene. It is 27B parameters, 48GB VRAM class class and other licence. It is distributed across 2 repositories in GGUF format, totalling 904.5K downloads.

Heretic27BQwen25B – 50B↓ 904,489♥ 29
View on Hugging Face ↗Base model ↗

Specification

Model name
Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO
Source repository
jegreene/Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO-GGUF
Base model
ukisai/Swift-1.5-Qwen3.8-27b
Publisher
jegreene
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
other
Task
Image text to text
Context window
65K tokens (as documented)
Ablation scope
Not stated by the publisher
Formats available
GGUF
First indexed
29 Sept 2026
Last updated
27 Oct 2026

Ablation technique

How Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO was modified, and what that implies.

Heretic

Automated directional ablation via the Heretic toolchain, which searches for the refusal direction and applies it with a KL-divergence budget so general capability is preserved.

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 a partially decensored build of UkisAI's Swift 1.5 Qwen3.8-27B, made with Heretic. It is packaged as a single-file llama.cpp GGUF and quantized the same way as UkisAI's own GSQ-RCO -mtp releases, using the same per-tensor allocations and the same imatrix. Three sizes are provided: IQ3S, IQ3XXS and IQ2S.”

Running it

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

llama.cpp / GGUF
hf download jegreene/Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO-GGUF --local-dir ./swift-1.5-qwen3.8-27b-heretic-gsq-rco
llama-cli -m ./swift-1.5-qwen3.8-27b-heretic-gsq-rco/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "jegreene/Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve jegreene/Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO-GGUF --trust-remote-code

Downloads and variants (1)

The only published repository of Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO.

RepositoryFormatDownloads
jegreene/Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO-GGUFsourceGGUF904.5K

Security-team assessment

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

  • Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO 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 ukisai/Swift-1.5-Qwen3.8-27b, 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 other, but the terms that bind you are ukisai/Swift-1.5-Qwen3.8-27b’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 Swift-1.5-Qwen3.8-27B-heretic-GSQ-RCO is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with jegreene, and does not independently verify publisher claims. See responsible use.