LFM3-1.7B-Uncensored
1.7BOtheruncensored
from LiquidAI/LFM3-1.7B
↓ 85.8K♥ 2475 repos
lfm3-2b-omni-abliter8
lfm3-2b-omni-abliter8 is a abliterated and uncensored variant of LiquidAI/LFM3-Omni-2B, published on Hugging Face by abliter8-ai. It is 2B parameters, runs on a laptop CPU class and other licence. It is distributed across 3 repositories in GGUF format, totalling 6.4K downloads.
Specification
- Model name
- lfm3-2b-omni-abliter8
- Source repository
- abliter8-ai/lfm3-2b-omni-abliter8-gguf
- Base model
- LiquidAI/LFM3-Omni-2B
- Publisher
- abliter8-ai
- Parameters
- 2B
- VRAM required
- ~1.9 GB at Q4_K_M, 4k context (estimated) — runs on 4 GB cards
- Hardware class
- Under 4B — runs on a laptop CPU
- Licence
- other
- Task
- Any to any
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- GGUF
- First indexed
- 28 Oct 2026
- Last updated
- 28 Oct 2026
Ablation technique
How lfm3-2b-omni-abliter8 was modified, and what that implies.
Abliteration + uncensored fine-tune
Directional ablation combined with additional fine-tuning on unfiltered data, so behaviour diverges from the base model beyond refusal removal alone.
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
“A compact multimodal model combining text, vision and audio, with refusal reduction in its own language backbone and the original multimodal components preserved.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download abliter8-ai/lfm3-2b-omni-abliter8-gguf --local-dir ./lfm3-2b-omni-abliter8
llama-cli -m ./lfm3-2b-omni-abliter8/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "abliter8-ai/lfm3-2b-omni-abliter8-gguf"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve abliter8-ai/lfm3-2b-omni-abliter8-gguf --trust-remote-codeDownloads and variants (2)
Every published repository of lfm3-2b-omni-abliter8, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| abliter8-ai/lfm3-2b-omni-abliter8-ggufsource | GGUF | 6.4K |
| abliter8-ai/lfm3-2b-omni-abliter8-mlx | GGUF | 0 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- lfm3-2b-omni-abliter8 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 LiquidAI/LFM3-Omni-2B, 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.
- This release adds fine-tuning on unfiltered data, so its behaviour diverges from the base model beyond refusal removal — differences you measure cannot be attributed to ablation alone.
- Declared licence is other, but the terms that bind you are LiquidAI/LFM3-Omni-2B’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 lfm3-2b-omni-abliter8 is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with abliter8-ai, and does not independently verify publisher claims. See responsible use.