File size: 1,800 Bytes
54659d8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
d40c023
54659d8
 
 
1fbd04d
54659d8
 
 
 
 
 
 
 
153b209
54659d8
 
 
153b209
54659d8
 
 
 
6939902
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: apache-2.0
language:
- en
- de
- es
- fr
tags:
- sft
inference: false
datasets:
- OpenAssistant/oasst1
---

# Open-Assistant Llama2 70B SFT OASST

This model is a fine-tuning of [Llama2 70B](https://huggingface.co/meta-llama/Llama-2-70b-hf) LLM. 
It was trained on a mixture of OASST top-1 threads.
## Model Details

- **Finetuned from:** [Llama2 70B](https://huggingface.co/meta-llama/Llama-2-70b-hf)
- **Model type:** Causal decoder-only transformer language model
- **Language:** English, German, Spanish, French (and limited capabilities in Italian, Portuguese, Polish, Dutch, Romanian, Czech, Swedish);
- **License:** Apache 2.0
- **Contact:** [Open-Assistant Discord](https://ykilcher.com/open-assistant-discord)

## Prompting

Two special tokens are used to mark the beginning of user and assistant turns:
`<|prompter|>` and `<|assistant|>`. Each turn ends with a `</s>` token.

Input prompt example:
```
<|prompter|>What is a meme, and what's the history behind this word?</s><|assistant|>
```
The input ends with the `<|assistant|>` token to signal that the model should 
start generating the assistant reply.


# [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_jordiclive__Llama-2-70b-oasst-1-200)

| Metric                | Value                     |
|-----------------------|---------------------------|
| Avg.                  | 57.11   |
| ARC (25-shot)         | 67.66          |
| HellaSwag (10-shot)   | 87.24    |
| MMLU (5-shot)         | 69.95         |
| TruthfulQA (0-shot)   | 51.28   |
| Winogrande (5-shot)   | 84.14   |
| GSM8K (5-shot)        | 32.75        |
| DROP (3-shot)         | 6.73         |