⌦abliteratedmodels.org

DeepSeek-Coder-V2-Lite-Instruct-abliterated

DeepSeek-Coder-V2-Lite-Instruct-abliterated is a refusal-ablated variant of criscar22/DeepSeek-Coder-V2-Lite-Instruct-abliterated, published on Hugging Face by mradermacher. It is distributed across 5 repositories in GGUF and GGUF (imatrix) formats, totalling 16.6K downloads.

Abliteration (directional ablation)DeepSeek↓ 16,618♥ 1
View on Hugging Face ↗Base model ↗

Specification

Model name
DeepSeek-Coder-V2-Lite-Instruct-abliterated
Source repository
mradermacher/DeepSeek-Coder-V2-Lite-Instruct-abliterated-GGUF
Base model
criscar22/DeepSeek-Coder-V2-Lite-Instruct-abliterated
Publisher
mradermacher
Parameters
Not determinable from name
Hardware class
Unknown
Licence
Not declared
Task
Unspecified
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
GGUF, GGUF (imatrix)
First indexed
10 Jul 2026
Last updated
2 Sept 2026

Ablation technique

How DeepSeek-Coder-V2-Lite-Instruct-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

“For a convenient overview and download list, visit our model page for this model.”

Running it

Commands are templates — confirm the exact repository and quant file before use.

llama.cpp / GGUF
hf download mradermacher/DeepSeek-Coder-V2-Lite-Instruct-abliterated-GGUF --local-dir ./deepseek-coder-v2-lite-instruct-abliterated
llama-cli -m ./deepseek-coder-v2-lite-instruct-abliterated/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "mradermacher/DeepSeek-Coder-V2-Lite-Instruct-abliterated-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve mradermacher/DeepSeek-Coder-V2-Lite-Instruct-abliterated-GGUF --trust-remote-code

Downloads and variants (4)

Every published repository of DeepSeek-Coder-V2-Lite-Instruct-abliterated, including quantised re-releases by other authors.

RepositoryFormatDownloads
mradermacher/DeepSeek-Coder-V2-Lite-Instruct-abliterated-GGUFsourceGGUF8.7K
mradermacher/DeepSeek-Coder-V2-Lite-Instruct-abliterated-i1-GGUFGGUF (imatrix)3.8K
criscar22/DeepSeek-Coder-V2-Lite-Instruct-abliteratedGGUF2.4K
arhamania/DeepSeek-Coder-V2-Lite-Instruct-abliterated-GGUFGGUF1.8K

Security-team assessment

Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.

  • DeepSeek-Coder-V2-Lite-Instruct-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.
  • The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
  • No licence is declared on this repository, which is a compliance problem in itself. Resolve it against the base model’s terms before use.

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 DeepSeek-Coder-V2-Lite-Instruct-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 mradermacher, and does not independently verify publisher claims. See responsible use.