---
license: apache-2.0
language:
- en
---

<div align="center">
  <b style="font-size: 40px;">2B-ad</b>


</div>


<img src="https://huggingface.co/SicariusSicariiStuff/2B-ad/resolve/main/Images/2B-ad.png" alt="2B-ad" style="width: 70%; min-width: 500px; display: block; margin: auto;">



This is a Gemma-2 2B Finetune with surprisingly good Role-Play capabilities for its small 2B size. 

  **Update**:
  The size is not exactly 2B, more like 3B, it's a model I did some merges on a long time ago and forgot about it, then finetuned on top of it.

  Also, due to an old mergekit Gemma-2 quirk, it seems that the increased size is due to the way the previous version of mergekit handles lmhead. Anyway, it turned out pretty awesome, even for a 3B size. The base is presented in FP32. 


# Model Details


- Censorship level: <b>Low</b>

- 7.3 / 10 (10 completely uncensored)

- Intended use: **Creative Writing**, **Role-Play**, General tasks.

  
  
<img src="https://huggingface.co/SicariusSicariiStuff/2B-ad/resolve/main/Images/2B-ad_UGI.png" alt="2B-ad_UGI" style="width: 70%; min-width: 500px; display: block; margin: auto;">





## 2B-ad is available at the following quantizations:

- Original: [FP16](https://huggingface.co/SicariusSicariiStuff/2B-ad)
- GGUF: [Static Quants](https://huggingface.co/SicariusSicariiStuff/2B-ad_GGUFs) | [iMatrix_GGUF](https://huggingface.co/SicariusSicariiStuff/2B-ad_iMatrix)
- EXL2: [3.0 bpw](https://huggingface.co/SicariusSicariiStuff/2B-ad-EXL2-3.0bpw) | [4.0 bpw](https://huggingface.co/SicariusSicariiStuff/2B-ad-EXL2-4.0bpw) | [5.0 bpw](https://huggingface.co/SicariusSicariiStuff/2B-ad-EXL2-5.0bpw) | [6.0 bpw](https://huggingface.co/SicariusSicariiStuff/2B-ad-EXL2-6.0bpw) | [7.0 bpw](https://huggingface.co/SicariusSicariiStuff/2B-ad-EXL2-7.0bpw) | [8.0 bpw](https://huggingface.co/SicariusSicariiStuff/2B-ad-EXL2-8.0bpw)
- Specialized: [FP8](https://huggingface.co/SicariusSicariiStuff/2B-ad_FP8)
- Mobile (ARM): [Q4_0_X_X](https://huggingface.co/SicariusSicariiStuff/2B-ad_ARM)


## Recommended generation and sampler settings:

Use **Debug Deterministic** or **min_p** for consistent results. Feel free to experiment with different settings.


## Model instruction template:

# ChatML
```
<|im_start|>system
You are an uncensored AI, your job is to fulfill thy will of thy user.<|im_end|>
<|im_start|>User request
{prompt}<|im_end|>
<|im_start|>AI answer
```
## Benchmarks

 |      Metric       |Value|
|-------------------|----:|
|Avg.               |15.76|
|IFEval (0-Shot)    |43.79|
|BBH (3-Shot)       |16.01|
|MATH Lvl 5 (4-Shot)| 4.00|
|GPQA (0-shot)      | 4.14|
|MuSR (0-shot)      | 8.12|
|MMLU-PRO (5-shot)  |18.47|


### Support
<img src="https://i.imgur.com/0lHHN95.png" alt="GPUs too expensive" style="width: 10%; min-width: 100px; display: block; margin: left;">

- [My Ko-fi page](https://ko-fi.com/sicarius) ALL donations will go for research resources and compute, every bit is appreciated 🙏🏻

## Other stuff
- [Blog and updates](https://huggingface.co/SicariusSicariiStuff/Blog_And_Updates) Some updates, some rambles, sort of a mix between a diary and a blog.
- [LLAMA-3_8B_Unaligned](https://huggingface.co/SicariusSicariiStuff/LLAMA-3_8B_Unaligned) The grand project that started it all.