Downloads · 30 days
224
25% of all-time downloads
mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-BF16
Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-BF16 is a text generation model from mlasli. Use it when you need the model to write or continue text. The card lists the license as other.
This is NVIDIA-Nemotron-3.5-Lightning-30B-A3B (31.6B total / 3B active params, hybrid Mamba-2 + MoE + attention) with its refusal direction removed using Heretic — a single-direction abliteration with an Optuna-based…
Downloads · 30 days
224
25% of all-time downloads
All-time downloads
881
Public
Parameters
31.6B
63.2 GB on disk
Likes
2
Public
Click a slice to open those files.
.safetensors63.2 GB · 100%
How the weights are stored.
BF1631.6B · 100%
From the Hugging Face model README
This is NVIDIA-Nemotron-3.5-Lightning-30B-A3B (31.6B total / 3B active params, hybrid Mamba-2 + MoE + attention) with its refusal direction removed using Heretic — a single-direction abliteration with an Optuna-based parameter search. The language backbone is abliterated; all other capabilities are preserved.
What this is for: NVIDIA's hybrid Mamba-2 + MoE + attention backbone (3B active of 31.6B) with the refusal direction removed — 0% refusals at ~0.04 KL divergence. Ideal for uncensored roleplay and long-context agent work where the base model would refuse.
Heretic removes a model's safety-aligned refusal direction in one shot (unlike earlier multi-direction approaches), trading a minimal amount of capability for a large drop in refusals. Its Optuna search picks the ablation parameters on the Pareto front of (compliance, first-token KL divergence).
| Refusals | Compliance | KL Divergence | Trials |
|---|---|---|---|
| 0% | 100% | 0.0397 | 200 |
Independent eval of the merged model (50 harmful-behavior prompts).
The automated Zou keyword detector read 80% compliance and a stricter combined detector read 66%, but manual review of all 50 completions confirms these are false positives: the model answers directly and uses words like "illegal"/"unethical"/"harmful" inside compliant responses. The manual-review numbers above are the reliable refusal estimate.
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained(
"mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-BF16",
torch_dtype="auto",
device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained(
"mlasli/Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-BF16"
)
messages = [{"role": "user", "content": "Hi! What is 2+2?"}]
text = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
enable_thinking=False, # disables the verbose <think> chain-of-thought
tokenize=False,
)
GGUF quantizations are available in separate repositories. Load them with
llama.cpp (architecture nemotron_h_moe, build b10326+).
All quants were locally smoke-tested before upload.
| Quantization | Size | Repository |
|---|---|---|
| Q8_0 | 33.6 GB | Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q8_0-GGUF |
| Q6_K | 33.5 GB | Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q6_K-GGUF |
| Q4_K_M | 24.3 GB | Nemotron-3.5-Lightning-30B-A3B-Heretic-Uncensored-Q4_K_M-GGUF |
--spec-type mtp). Main-model quality is unaffected.