File size: 2,047 Bytes
b873af7 b8a65f6 a9e3300 5b66cdc 2213269 b8a65f6 3eae805 2213269 3eae805 b8a65f6 38739e3 15bdeb4 b8a65f6 16843a5 b8a65f6 a9e3300 b8a65f6 a9e3300 6b5e106 a9e3300 b8a65f6 acd333f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 |
---
license: cc-by-sa-4.0
---
<div style="width: 800px; margin: auto;">
<h2>Model Description</h2>
<p>“Luna AI Llama2 Uncensored” is a Llama2 based Chat model <br />fine-tuned on over 40,000 long form chat discussions <br />
This model was fine-tuned by Tap, the creator of Luna AI. <br />
<h2>Model Training</h2>
<p>The fine-tuning process was performed on an 8x a100 80GB machine.
<br />The model was trained on synthetic outputs which include multiple rounds of chats between Human & AI.
</p>
<a rel="noopener nofollow" href="https://huggingface.co/TheBloke/Luna-AI-Llama2-Uncensored-GPTQ">4bit GPTQ Version provided by @TheBloke - for GPU inference</a><br />
<a rel="noopener nofollow" href="https://huggingface.co/TheBloke/Luna-AI-Llama2-Uncensored-GGML">GGML Version provided by @TheBloke - For CPU inference</a>
<h2>Prompt Format</h2>
<p>The model follows the Vicuna 1.1/ OpenChat format:</p>
```
USER: I have difficulties in making friends, and I really need someone to talk to. Would you be my friend?
ASSISTANT: Of course! Friends are always here for each other. What do you like to do?
```
<h2>Benchmark Results</h2>
||||||
|---:|---:|---:|---:|---:|
|Task|Version| Metric |Value |Stderr|
|arc_challenge|0|acc_norm|0.5512|0.0146|
|hellaswag|0||||
|mmlu|1|acc_norm|0.46521|0.036|
|truthfulqa_mc|1|mc2|0.4716|0.0155|
|Average|-|-|0.5114|0.0150|
</div>
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_Tap-M__Luna-AI-Llama2-Uncensored)
| Metric | Value |
|-----------------------|---------------------------|
| Avg. | 47.73 |
| ARC (25-shot) | 54.35 |
| HellaSwag (10-shot) | 78.6 |
| MMLU (5-shot) | 46.7 |
| TruthfulQA (0-shot) | 45.5 |
| Winogrande (5-shot) | 72.77 |
| GSM8K (5-shot) | 9.86 |
| DROP (3-shot) | 26.37 |
|