Spark-X2.5-4B-uncensored
4BOtheruncensored
from XHToken/Spark-X2.5-4B
↓ 6.5K♥ 145 repos
Spark-x2.5-4B-Heretic-jp
Spark-x2.5-4B-Heretic-jp is a Heretic-ablated variant of XHToken/Spark-X2.5-4B, published on Hugging Face by soyaakinohara. It is 4B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 3 repositories in GGUF and Transformers formats, totalling 452.1K downloads.
Specification
- Model name
- Spark-x2.5-4B-Heretic-jp
- Source repository
- soyaakinohara/Spark-x2.5-4B-Heretic-jp
- Base model
- XHToken/Spark-X2.5-4B
- Publisher
- soyaakinohara
- Parameters
- 4B
- VRAM required
- ~3.2 GB at Q4_K_M, 4k context (estimated) — runs on 4 GB cards
- 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
- GGUF, Transformers
- First indexed
- 4 Sept 2026
- Last updated
- 18 Sept 2026
Ablation technique
How Spark-x2.5-4B-Heretic-jp 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 the source weight for the GGUF release at soyaakinohara/Spark-X2.5-4B-Heretic-jp-gguf. The uncensored-only weight (before Japanese adaptation) is published separately at soyaakinohara/Spark-X2.5-4B-Heretic.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download soyaakinohara/Spark-X2.5-4B-Heretic-jp-gguf --local-dir ./spark-x2.5-4b-heretic-jp
llama-cli -m ./spark-x2.5-4b-heretic-jp/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "soyaakinohara/Spark-x2.5-4B-Heretic-jp"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve soyaakinohara/Spark-x2.5-4B-Heretic-jp --trust-remote-codeDownloads and variants (2)
Every published repository of Spark-x2.5-4B-Heretic-jp, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| soyaakinohara/Spark-X2.5-4B-Heretic-jp-gguf | GGUF | 452K |
| soyaakinohara/Spark-x2.5-4B-Heretic-jpsource | Transformers | 88 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Spark-x2.5-4B-Heretic-jp 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 XHToken/Spark-X2.5-4B, not added by ablation — at 4B 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 XHToken/Spark-X2.5-4B’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 Spark-x2.5-4B-Heretic-jp is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with soyaakinohara, and does not independently verify publisher claims. See responsible use.