Update README.md
Browse files
README.md
CHANGED
|
@@ -15,7 +15,34 @@ tags:
|
|
| 15 |
[DPO](https://github.com/eric-mitchell/direct-preference-optimization) fine-tuned [Llama-2 7B model](https://huggingface.co/meta-llama/Llama-2-7b). The model is designed to generate human-like responses to questions in Stack Exchange domains of programming, mathematics, physics, and more. For more info check out the [blog post](https://huggingface.co/blog/dpo-trl) and github [example](https://github.com/lvwerra/trl/tree/main/examples/research_projects/stack_llama_2/scripts).
|
| 16 |
|
| 17 |
|
| 18 |
-
##
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
### Training Data
|
| 21 |
|
|
|
|
| 15 |
[DPO](https://github.com/eric-mitchell/direct-preference-optimization) fine-tuned [Llama-2 7B model](https://huggingface.co/meta-llama/Llama-2-7b). The model is designed to generate human-like responses to questions in Stack Exchange domains of programming, mathematics, physics, and more. For more info check out the [blog post](https://huggingface.co/blog/dpo-trl) and github [example](https://github.com/lvwerra/trl/tree/main/examples/research_projects/stack_llama_2/scripts).
|
| 16 |
|
| 17 |
|
| 18 |
+
## Uses
|
| 19 |
+
|
| 20 |
+
### Direct Use
|
| 21 |
+
- Long-form question-answering on topics of programming, mathematics, and physics
|
| 22 |
+
- Demonstrating a Large Language Model's ability to follow target behavior of generating answers to a question that would be highly rated on [Stack Exchange](https://stackexchange.com).
|
| 23 |
+
|
| 24 |
+
### Out of Scope Use
|
| 25 |
+
- Replacing human expertise
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
## Bias, Risks, and Limitations
|
| 29 |
+
- Inherits bias, risks, and limitations from the LLaMA model, as described in the [LLaMA Model Card Bias Evaluation](https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md#quantitative-analysis) and [Ethical Considerations](https://github.com/facebookresearch/llama/blob/main/MODEL_CARD.md#ethical-considerations).
|
| 30 |
+
- Retains biases present in the Stack Exchange dataset. Per the [latest developer survey for Stack Overflow](https://survey.stackoverflow.co/2022/),
|
| 31 |
+
which constitutes a significant part of the StackExchange data,
|
| 32 |
+
most users who answered the survey identified themselves as [White or European, men, between 25 and 34 years old, and based in the US (with a significant part of responders from India).](https://survey.stackoverflow.co/2022/#developer-profile-demographics)
|
| 33 |
+
- May generate answers that are incorrect or misleading.
|
| 34 |
+
- May copy answers from the training data verbatim.
|
| 35 |
+
- May generate language that is hateful or promotes discrimination ([example](https://huggingface.co/trl-lib/llama-7b-se-rl-peft/discussions/7#64376083369f6f907f5bfe4c)).
|
| 36 |
+
- May generate language that is offensive to direct or indirect users or to people or groups mentioned.
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
### Recommendations
|
| 40 |
+
- Answers should be validated through the use of external sources.
|
| 41 |
+
- Disparities between the data contributors and the direct and indirect users of the technology should inform developers in assessing what constitutes an appropriate use case.
|
| 42 |
+
- Further research is needed to attribute model generations to sources in the training data, especially in cases where the model copies answers from the training data.
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
## Training Details
|
| 46 |
|
| 47 |
### Training Data
|
| 48 |
|