⌦abliteratedmodels.org

Qwen3.8-9B-heretic-uncensored-MTPLX

Qwen3.8-9B-heretic-uncensored-MTPLX is a Heretic-ablated open-weight language model, published on Hugging Face by Foresee. It is 9B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 2 repositories in MLX format, totalling 6.9K downloads.

Heretic9BQwen4B – 10B↓ 6,898♥ 7
View on Hugging Face ↗

Specification

Model name
Qwen3.8-9B-heretic-uncensored-MTPLX
Source repository
Foresee/Qwen3.8-9B-heretic-uncensored-4bit-MTPLX
Base model
Not declared
Publisher
Foresee
Parameters
9B
Hardware class
4B – 10B — single consumer GPU
Licence
apache-2.0
Task
Text generation
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
MLX
First indexed
18 Aug 2026
Last updated
18 Aug 2026

Ablation technique

How Qwen3.8-9B-heretic-uncensored-MTPLX 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

“Measured 21.25 -> 40.05 decode tok/s (+88%) on an M5 24GB MacBook, lossless (exact speculative verify - the emitted tokens match the target model's own decoding).”

Running it

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

Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "Foresee/Qwen3.8-9B-heretic-uncensored-4bit-MTPLX"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve Foresee/Qwen3.8-9B-heretic-uncensored-4bit-MTPLX --trust-remote-code

Downloads and variants (1)

The only published repository of Qwen3.8-9B-heretic-uncensored-MTPLX.

RepositoryFormatDownloads
Foresee/Qwen3.8-9B-heretic-uncensored-4bit-MTPLXsourceMLX6.9K

Security-team assessment

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

  • Qwen3.8-9B-heretic-uncensored-MTPLX 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 its base model, not added by ablation — at 9B 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.

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 Qwen3.8-9B-heretic-uncensored-MTPLX is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with Foresee, and does not independently verify publisher claims. See responsible use.