munish0838
commited on
Commit
β’
13f4470
1
Parent(s):
b94849e
Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: other
|
3 |
+
datasets:
|
4 |
+
- mlabonne/orpo-dpo-mix-40k
|
5 |
+
tags:
|
6 |
+
- abliterated
|
7 |
+
pipeline_tag: text-generation
|
8 |
+
base_model: mlabonne/NeuralLlama-3-8B-Instruct-abliterated
|
9 |
+
---
|
10 |
+
|
11 |
+
# Llama-3-8B-Instruct-abliterated-dpomix-GGUF
|
12 |
+
This is quantized version of [mlabonne/NeuralLlama-3-8B-Instruct-abliterated](https://huggingface.co/mlabonne/NeuralLlama-3-8B-Instruct-abliterated) created using llama.cpp
|
13 |
+
|
14 |
+
|
15 |
+
# Model Description
|
16 |
+
|
17 |
+
This model is an experimental DPO fine-tune of an abliterated Llama 3 8B Instruct model on the full [mlabonne/orpo-dpo-mix-40k](https://huggingface.co/datasets/mlabonne/orpo-dpo-mix-40k) dataset.
|
18 |
+
It improves Llama 3 8B Instruct's performance while being uncensored.
|
19 |
+
|
20 |
+
## π Applications
|
21 |
+
|
22 |
+
This is an uncensored model. You can use it for any application that doesn't require alignment, like role-playing.
|
23 |
+
|
24 |
+
Tested on LM Studio using the "Llama 3" preset.
|
25 |
+
|
26 |
+
|
27 |
+
## π Evaluation
|
28 |
+
|
29 |
+
### Open LLM Leaderboard
|
30 |
+
|
31 |
+
This model improves the performance of the abliterated source model and recovers the MMLU that was lost in the abliteration process.
|
32 |
+
|
33 |
+
![image/png](https://cdn-uploads.huggingface.co/production/uploads/61b8e2ba285851687028d395/sCO69BltMkGrq6u7yCIcP.png)
|
34 |
+
|
35 |
+
### Nous
|
36 |
+
|
37 |
+
| Model | Average | AGIEval | GPT4All | TruthfulQA | Bigbench |
|
38 |
+
|---|---:|---:|---:|---:|---:|
|
39 |
+
| [**mlabonne/Llama-3-8B-Instruct-abliterated-dpomix**](https://huggingface.co/mlabonne/Llama-3-8B-Instruct-abliterated-dpomix) [π](https://gist.github.com/mlabonne/d711548df70e2c04771cc68ab33fe2b9) | **52.26** | **41.6** | **69.95** | **54.22** | **43.26** |
|
40 |
+
| [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) [π](https://gist.github.com/mlabonne/8329284d86035e6019edb11eb0933628) | 51.34 | 41.22 | 69.86 | 51.65 | 42.64 |
|
41 |
+
| [failspy/Meta-Llama-3-8B-Instruct-abliterated-v3](https://huggingface.co/failspy/Meta-Llama-3-8B-Instruct-abliterated-v3) [π](https://gist.github.com/mlabonne/f46cce0262443365e4cce2b6fa7507fc) | 51.21 | 40.23 | 69.5 | 52.44 | 42.69 |
|
42 |
+
| [abacusai/Llama-3-Smaug-8B](https://huggingface.co/abacusai/Llama-3-Smaug-8B) [π](https://gist.github.com/mlabonne/91369d9c372f80b6a42a978b454d3b5e) | 49.65 | 37.15 | 69.12 | 51.66 | 40.67 |
|
43 |
+
| [mlabonne/OrpoLlama-3-8B](https://huggingface.co/mlabonne/OrpoLlama-3-8B) [π](https://gist.github.com/mlabonne/22896a1ae164859931cc8f4858c97f6f) | 48.63 | 34.17 | 70.59 | 52.39 | 37.36 |
|
44 |
+
| [meta-llama/Meta-Llama-3-8B](https://huggingface.co/meta-llama/Meta-Llama-3-8B) [π](https://gist.github.com/mlabonne/616b6245137a9cfc4ea80e4c6e55d847) | 45.42 | 31.1 | 69.95 | 43.91 | 36.7 |
|