LL

Impish_LLAMA_4B_Abliterated

LlamaGGUF by mradermacherUpdated 2026-01-28
Download on Hugging Face
Parameters4.51B
Context128K
Layers32
Architecturellama
Smallest quant0.0 GB
Quick answer

Impish_LLAMA_4B_Abliterated needs about 4.4 GB of memory at Q4_K_M (a 2.8 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 25 quants, from 0.0 GB to 3.7 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.

Impish_LLAMA_4B_Abliterated – a 4.51B Llama model with removed safety guardrails

Impish_LLAMA_4B_Abliterated is a 4.51B parameter Llama model that has had its refusal mechanisms surgically removed. It retains the full capabilities of the original while eliminating safety guardrails through orthogonalization techniques. The model is available in multiple GGUF quantizations.

Good at

  • Large context handling
  • High parameter count
  • Versatile language generation

Limits

  • No safety guardrails
  • No vision support
  • Requires 8k context for memory estimates

Best for

  • Research on alignment
  • Testing model behavior
  • Educational demonstrations

Who it’s for. Researchers and developers who need a fully capable Llama model without built‑in safety constraints. Use with caution and in controlled environments.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.0 GB 1.6 GB
Fits Get file →
IQ1_S1.2 GB 2.8 GB
Fits Get file →
IQ1_M1.3 GB 2.9 GB
Fits Get file →
IQ2_XXS1.4 GB 3.0 GB
Fits Get file →
IQ2_XS1.5 GB 3.1 GB
Fits Get file →
IQ2_S1.6 GB 3.2 GB
Fits Get file →
IQ2_M1.7 GB 3.3 GB
Fits Get file →
Q2_K_S1.7 GB 3.3 GB
Fits Get file →
Q2_K1.8 GB 3.4 GB
Fits Get file →
IQ3_XXS1.9 GB 3.5 GB
Fits Get file →
IQ3_XS2.0 GB 3.6 GB
Fits Get file →
Q3_K_S2.1 GB 3.7 GB
Fits Get file →
IQ3_S2.1 GB 3.7 GB
Fits Get file →
IQ3_M2.2 GB 3.8 GB
Fits Get file →
Q3_K_M2.3 GB 3.9 GB
Fits Get file →
Q3_K_L2.5 GB 4.0 GB
Fits Get file →
IQ4_XS2.5 GB 4.1 GB
Fits Get file →
Q4_02.7 GB 4.2 GB
Fits Get file →
IQ4_NL2.7 GB 4.2 GB
Fits Get file →
Q4_K_S2.7 GB 4.2 GB
Fits Get file →
Q4_K_M2.8 GB 4.4 GB
Fits Get file →
Q4_12.9 GB 4.5 GB
Fits Get file →
Q5_K_S3.2 GB 4.7 GB
Fits Get file →
Q5_K_M3.2 GB 4.8 GB
Fits Get file →
Q6_K3.7 GB 5.3 GB
Fits Get file →

For 8 GB VRAM, the Q4_K_M quant (~4.4 GB memory at 8k context) is suitable. If more memory is available, Q5_K_M (~4.8 GB) or Q6_K (~5.3 GB) can be used. The smallest quant, IQ1_S, uses only 2.8 GB memory at 8k context and is best for low‑budget setups.

What GPU do you need for Impish_LLAMA_4B_Abliterated?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is Q3_K_L (2.5 GB file, about 4.0 GB in use).
  • 6 GB VRAM: the best quant that fits is Q6_K (3.7 GB file, about 5.3 GB in use).

Which Impish_LLAMA_4B_Abliterated quant should you choose?

Q4_K_M is the usual starting point for Impish_LLAMA_4B_Abliterated: it balances file size and quality well. Here it is a 2.8 GB file.

If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters4.51B
Layers32
Trained context131,072 tokens
KV cache (fp16)128 KB per token

Source

Quantized bymradermacher
FamilyLlama
Licensellama3.1
Downloads (HF)515
Last updated2026-01-28

Frequently asked questions about Impish_LLAMA_4B_Abliterated

How much VRAM does Impish_LLAMA_4B_Abliterated need?

About 4.4 GB at Q4_K_M with an 8K context window. That is the 2.8 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Impish_LLAMA_4B_Abliterated on an 8 GB GPU?

Yes, with Q6_K: the 3.7 GB file needs about 5.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Impish_LLAMA_4B_Abliterated GGUF?

It depends on the quant: file is 0.0 GB, IQ1_S is 1.2 GB, IQ1_M is 1.3 GB, IQ2_XXS is 1.4 GB, IQ2_XS is 1.5 GB, IQ2_S is 1.6 GB and 19 more.

What context length does Impish_LLAMA_4B_Abliterated support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.

Where can I download Impish_LLAMA_4B_Abliterated GGUF?

The files are on Hugging Face in the mradermacher/Impish_LLAMA_4B_Abliterated-i1-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Impish_LLAMA_4B_Abliterated use?

Hugging Face lists the license as llama3.1. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Impish_LLAMA_4B_Abliterated on 8 GB VRAM?

Yes, the Q4_K_M quant uses approximately 4.4 GB memory at 8k context, fitting within an 8 GB VRAM budget.

Does this model support vision tasks?

No, the model has no vision capabilities.

What license applies to this model?

The model is licensed under llama3.1.

Other GGUFs of Impish_LLAMA_4B_Abliterated

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Llama models

All Llama →

Other small models

All small models →