⌦abliteratedmodels.org

Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored

Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored is a Heretic-ablated variant of nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16, published on Hugging Face by mlasli. It is 30B total parameters (mixture-of-experts), 48GB VRAM class class and other licence. It is distributed across 5 repositories in Full precision and GGUF formats, totalling 6K downloads.

Heretic30B MoENemotron25B – 50B↓ 6,034♥ 3
View on Hugging Face ↗Base model ↗

Specification

Model name
Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored
Source repository
mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q4_K_M-GGUF
Base model
nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16
Publisher
mlasli
Parameters
30B (mixture-of-experts)
VRAM required
~19 GB at Q4_K_M, 4k context (estimated) — runs on 24 GB cards
Hardware class
25B – 50B — 48GB VRAM class
Licence
other
Task
Text generation
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
Full precision, GGUF
First indexed
16 Aug 2026
Last updated
17 Aug 2026

Ablation technique

How Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored 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

“What this is for: the same abliterated hybrid Mamba-MoE model, quantized to 24.3 GB so it runs locally via llama.cpp.”

Running it

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

Ollama (publisher-documented)
ollama run nemotron-3.5-30b-heretic-q4_k_m
llama.cpp / GGUF
hf download mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q4_K_M-GGUF --local-dir ./nemotron-3.5-lightning-30b-a3b-heretic-uncensored
llama-cli -m ./nemotron-3.5-lightning-30b-a3b-heretic-uncensored/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q4_K_M-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q4_K_M-GGUF --trust-remote-code

Downloads and variants (4)

Every published repository of Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored, including quantised re-releases by other authors.

RepositoryFormatDownloads
mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q4_K_M-GGUFsourceGGUF4.6K
mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q6_K-GGUFGGUF735
mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q8_0-GGUFGGUF466
mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-BF16Full precision190

Security-team assessment

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

  • Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored 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 nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16, not added by ablation — at 30B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
  • The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
  • Declared licence is other, but the terms that bind you are nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16’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 Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with mlasli, and does not independently verify publisher claims. See responsible use.