Spark-x2.5-4B-Heretic-jp
4BOtherheretic
from XHToken/Spark-X2.5-4B
↓ 452.1K♥ 182 repos
Ministral-3-8B-Reasoning-UNCENSORED-heretic
Ministral-3-8B-Reasoning-UNCENSORED-heretic is a Heretic-ablated variant of unsloth/Ministral-3-8B-Reasoning-2512, published on Hugging Face by xtools-at. It is 8B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 4 repositories in Transformers, GGUF and GGUF (imatrix) formats, totalling 6.8K downloads.
Specification
- Model name
- Ministral-3-8B-Reasoning-UNCENSORED-heretic
- Source repository
- xtools-at/Ministral-3-8B-Reasoning-UNCENSORED-heretic
- Base model
- unsloth/Ministral-3-8B-Reasoning-2512
- Publisher
- xtools-at
- 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
- apache-2.0
- Task
- Unspecified
- Context window
- 256K tokens (as documented)
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, GGUF (imatrix)
- First indexed
- 20 Oct 2026
- Last updated
- 22 Oct 2026
Ablation technique
How Ministral-3-8B-Reasoning-UNCENSORED-heretic 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
“A balanced model in the Ministral 3 family, Ministral 3 8B is a powerful, efficient tiny language model with vision capabilities.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download mradermacher/Ministral-3-8B-Reasoning-UNCENSORED-heretic-i1-GGUF --local-dir ./ministral-3-8b-reasoning-uncensored-heretic
llama-cli -m ./ministral-3-8b-reasoning-uncensored-heretic/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "xtools-at/Ministral-3-8B-Reasoning-UNCENSORED-heretic"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve xtools-at/Ministral-3-8B-Reasoning-UNCENSORED-heretic --trust-remote-codeDownloads and variants (3)
Every published repository of Ministral-3-8B-Reasoning-UNCENSORED-heretic, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/Ministral-3-8B-Reasoning-UNCENSORED-heretic-i1-GGUF | GGUF (imatrix) | 5.4K |
| mradermacher/Ministral-3-8B-Reasoning-UNCENSORED-heretic-GGUF | GGUF | 1.3K |
| xtools-at/Ministral-3-8B-Reasoning-UNCENSORED-hereticsource | Transformers | 45 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Ministral-3-8B-Reasoning-UNCENSORED-heretic 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 unsloth/Ministral-3-8B-Reasoning-2512, 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 apache-2.0, but the terms that bind you are unsloth/Ministral-3-8B-Reasoning-2512’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 Ministral-3-8B-Reasoning-UNCENSORED-heretic is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with xtools-at, and does not independently verify publisher claims. See responsible use.