Spark-x2.5-4B-Heretic-jp
4BOtherheretic
from XHToken/Spark-X2.5-4B
↓ 452.1K♥ 182 repos
WizardLM-2-7B-abliterated
WizardLM-2-7B-abliterated is a refusal-ablated variant of fearlessdots/WizardLM-2-7B-abliterated, published on Hugging Face by fearlessdots. It is 7B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 12 repositories in Transformers, GGUF and EXL2 formats, totalling 5.3K downloads.
Specification
- Model name
- WizardLM-2-7B-abliterated
- Source repository
- fearlessdots/WizardLM-2-7B-abliterated
- Base model
- fearlessdots/WizardLM-2-7B-abliterated
- Publisher
- fearlessdots
- Parameters
- 7B
- VRAM required
- ~5 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
- Text generation
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, EXL2
- First indexed
- 23 May 2024
- Last updated
- 27 Jan 2026
Ablation technique
How WizardLM-2-7B-abliterated was modified, and what that implies.
Abliteration (directional ablation)
The single residual-stream direction that mediates refusal is identified from harmful/harmless prompt pairs and projected out of the model weights.
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 WizardLM-2-7B model with orthogonalized bfloat16 safetensor weights, based on the implementation by @failspy. For more info:”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download mradermacher/WizardLM-2-7B-abliterated-GGUF --local-dir ./wizardlm-2-7b-abliterated
llama-cli -m ./wizardlm-2-7b-abliterated/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "fearlessdots/WizardLM-2-7B-abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve fearlessdots/WizardLM-2-7B-abliterated --trust-remote-codeDownloads and variants (11)
Every published repository of WizardLM-2-7B-abliterated, including quantised re-releases by other authors.
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- WizardLM-2-7B-abliterated 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 fearlessdots/WizardLM-2-7B-abliterated, not added by ablation — at 7B 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 fearlessdots/WizardLM-2-7B-abliterated’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 WizardLM-2-7B-abliterated is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with fearlessdots, and does not independently verify publisher claims. See responsible use.