andreaskoepf
commited on
Commit
•
464401b
1
Parent(s):
3af8dbf
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,56 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
---
|
4 |
+
|
5 |
+
# Open-Assistant SFT-1 12B Model
|
6 |
+
|
7 |
+
|
8 |
+
This is the first iteration English supervised-fine-tuning (SFT) model of
|
9 |
+
the [Open-Assistant](https://github.com/LAION-AI/Open-Assistant) project.
|
10 |
+
It is based on a Pythia 12B that was fine-tuned on ~22k human demonstrations
|
11 |
+
of assistant conversations collected through the
|
12 |
+
[https://open-assistant.io/](https://open-assistant.io/) human feedback web
|
13 |
+
app before March 7, 2023.
|
14 |
+
|
15 |
+
### Model Description
|
16 |
+
|
17 |
+
- **Developed by:** Open-Assistant Contributors
|
18 |
+
- **Model type:** Conversational
|
19 |
+
- **Language:** English
|
20 |
+
- **License:** Apache 2.0
|
21 |
+
- **Finetuned from:** [EleutherAI / pythia-12b-deduped](https://huggingface.co/EleutherAI/pythia-12b-deduped)
|
22 |
+
|
23 |
+
### Sources
|
24 |
+
|
25 |
+
- **Repository:** [Open-Assistant/model/model_training](https://github.com/LAION-AI/Open-Assistant/tree/main/model/model_training)
|
26 |
+
- **Demo:** View Assistant replies for 250 random prompts [here](https://open-assistant.github.io/oasst-model-eval/?f=https%3A%2F%2Fraw.githubusercontent.com%2FOpen-Assistant%2Foasst-model-eval%2Fmain%2Fsampling_reports%2Foasst-sft%2F2023-03-09_andreaskoepf_oasst-1_12b_7000_sampling_noprefix_lottery.json) ([sampling code](https://github.com/Open-Assistant/oasst-model-eval/blob/3d71f3be100c05cd8ddb568365e036a29fbff8c7/model_eval/manual/sampling_report.py)).
|
27 |
+
|
28 |
+
### Reply Generation
|
29 |
+
|
30 |
+
Two special tokens are used to mark the beginning of user and assistant turns:
|
31 |
+
`<|prompter|>` and `<|assistant|>`. Each turn ends with a `<|endoftext|>` token.
|
32 |
+
|
33 |
+
Input example:
|
34 |
+
```
|
35 |
+
<|prompter|>What is a meme, and what's the history behind this word?<|endoftext|><|assistant|>
|
36 |
+
```
|
37 |
+
|
38 |
+
Generated continuation:
|
39 |
+
```
|
40 |
+
A visual joke often accompanied by text that spreads rapidly from individual to individual. The exact origin of the term "meme" is unclear, but it was first used in the 1990s by Richard Dawkins in his book "The Selfish Gene." He defined a meme as an idea, behavior, or style that is transmitted from one person to another within a culture.
|
41 |
+
|
42 |
+
Dawkins' use of the term caught on among scholars and journalists, who began to apply the concept to other types of cultural items such as advertisements, fashion trends, and jokes. Over time, the term "meme" has come to be used more generally to describe any social behavior that can be repeated and altered by individuals. Today, the term is widely recognized and applied in fields ranging from anthropology to psychology to marketing.<|endoftext|>
|
43 |
+
```
|
44 |
+
|
45 |
+
## Limitations
|
46 |
+
|
47 |
+
The model is known to fail horribly at answering math and coding questions.
|
48 |
+
|
49 |
+
Beware of hallucinations: Outputs are often factually wrong or misleading. Replies might look convincing (at first glance) while containing completely made up false statements.
|
50 |
+
|
51 |
+
This model is usable only for English conversations.
|
52 |
+
|
53 |
+
## Contact the Authors
|
54 |
+
|
55 |
+
- Github: [LAION-AI / Open-Assistant](https://github.com/LAION-AI/Open-Assistant)
|
56 |
+
- Discord: [Open-Assistant](https://ykilcher.com/open-assistant-discord)
|