Qwen3.5-2B-Uncensored-HauhauCS-Aggressive
2BQwenuncensored
from Qwen/Qwen3.5-2B
↓ 80.7K♥ 13110 repos
Qwen3.8-2B-Heretic
Qwen3.8-2B-Heretic is a Heretic-ablated variant of empero-ai/Qwen3.8-2B-Distill, published on Hugging Face by MihaiPopa-1. It is 2B parameters, runs on a laptop CPU class and apache-2.0 licence. It is distributed across 5 repositories in Transformers, GGUF and GGUF (imatrix) formats, totalling 6K downloads.
Specification
- Model name
- Qwen3.8-2B-Heretic
- Source repository
- MihaiPopa-1/Qwen3.8-2B-Heretic-Max
- Base model
- empero-ai/Qwen3.8-2B-Distill
- Publisher
- MihaiPopa-1
- Parameters
- 2B
- Hardware class
- Under 4B — runs on a laptop CPU
- 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
- 23 Aug 2026
- Last updated
- 24 Aug 2026
Ablation technique
How Qwen3.8-2B-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
“Qwen3.8-2B is a full-parameter distillation of Qwen3.8 2.4T A95B into the Qwen3.5-2B architecture — the smallest member of the family, trained on the same curriculum as its larger siblings. The student saw ~30,000 curated teacher traces from our internal Qwen3.8 distillation datasets — dense chain-of-thought spanning...”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download mradermacher/Qwen3.8-2B-Heretic-Max-i1-GGUF --local-dir ./qwen3.8-2b-heretic
llama-cli -m ./qwen3.8-2b-heretic/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "MihaiPopa-1/Qwen3.8-2B-Heretic-Max"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve MihaiPopa-1/Qwen3.8-2B-Heretic-Max --trust-remote-codeDownloads and variants (4)
Every published repository of Qwen3.8-2B-Heretic, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/Qwen3.8-2B-Heretic-Max-i1-GGUF | GGUF (imatrix) | 3K |
| mradermacher/Qwen3.8-2B-Heretic-Max-GGUF | GGUF | 2.2K |
| aicoder43210/Qwen3.8-2B-Heretic-Max-Q4_K_M-GGUF | GGUF | 506 |
| MihaiPopa-1/Qwen3.8-2B-Heretic-Maxsource | Transformers | 314 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwen3.8-2B-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 empero-ai/Qwen3.8-2B-Distill, not added by ablation — at 2B 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 empero-ai/Qwen3.8-2B-Distill’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 Qwen3.8-2B-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 MihaiPopa-1, and does not independently verify publisher claims. See responsible use.