Qwythos-9B-v2
9BOtheruncensored
from empero-ai/Qwythos-9B-v2
↓ 639.9K♥ 26611 repos
Katarau-9B-ru-RP-nsfw
Katarau-9B-ru-RP-nsfw is a Heretic-ablated variant of katafiek/Katarau-9B-ru-RP-nsfw, published on Hugging Face by katafiek. It is 9B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 5 repositories in Transformers, GGUF and GGUF (imatrix) formats, totalling 9.8K downloads.
Specification
- Model name
- Katarau-9B-ru-RP-nsfw
- Source repository
- katafiek/Katarau-9B-ru-RP-nsfw
- Base model
- katafiek/Katarau-9B-ru-RP-nsfw
- Publisher
- katafiek
- 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
- Transformers, GGUF, GGUF (imatrix)
- First indexed
- 5 Jul 2026
- Last updated
- 16 Jul 2026
Ablation technique
How Katarau-9B-ru-RP-nsfw 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
“All sources were originally written in or translated to Russian with Deepseek-v4-pro/Deepseek-v4-flash. Only ~2% of the dataset contains blocks — so the model ships in no-think mode, and it's recommended to run it without thinking.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download mradermacher/Katarau-9B-ru-RP-nsfw-GGUF --local-dir ./katarau-9b-ru-rp-nsfw
llama-cli -m ./katarau-9b-ru-rp-nsfw/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "katafiek/Katarau-9B-ru-RP-nsfw"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve katafiek/Katarau-9B-ru-RP-nsfw --trust-remote-codeDownloads and variants (4)
Every published repository of Katarau-9B-ru-RP-nsfw, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/Katarau-9B-ru-RP-nsfw-GGUF | GGUF | 4.8K |
| katafiek/Katarau-9B-ru-RP-nsfw-GGUF | GGUF | 3K |
| mradermacher/Katarau-9B-ru-RP-nsfw-i1-GGUF | GGUF (imatrix) | 1.6K |
| katafiek/Katarau-9B-ru-RP-nsfwsource | Transformers | 317 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Katarau-9B-ru-RP-nsfw 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 katafiek/Katarau-9B-ru-RP-nsfw, 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.
- Declared licence is apache-2.0, but the terms that bind you are katafiek/Katarau-9B-ru-RP-nsfw’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 Katarau-9B-ru-RP-nsfw is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with katafiek, and does not independently verify publisher claims. See responsible use.