GitXplorerGitXplorer
j

toki-pona-orpo

public
1 stars
0 forks
0 issues

Commits

List of commits on branch main.
Verified
28c8935f5e0aaada75eb8617b0a8a468d0715b16

Update dataset.json

jjohnpaulbin committed 6 months ago
Verified
77b9e5aa12f31338cb996ba75f7cc615715c7c86

Update README.md

jjohnpaulbin committed 6 months ago
Verified
b97a634e8cb992ebb96da2eca364e8de2669373f

Create dataset.json

jjohnpaulbin committed 6 months ago
Verified
b032ee38f21035ee24decbbe77e4140635792659

Initial commit

jjohnpaulbin committed 6 months ago

README

The README file for this repository.

toki-pona-orpo

This project aims to provide a high-quality toki pona ORPO dataset, most suited for a chat bot LLM model.

Data collected here are from a private data collection web ui, where users can contribute and edit.

Please DM me on discord for more information & how to contribute: mega_b

Data formatting

The dataset is a JSON file. Each entry of the ORPO dataset contains the following:

  • Instruction - Can be either an instruction to the chatbot, or a question.
  • Input - If the instruction calls for the input, then this is filled out, other wise it's left blank.
  • Accepted - This the best response to the instruction. (e.g following the instruction, being nice, 100% accurate response.)
  • Rejected - This is an example of a bad response to the instruction. (e.g inaccurate toki pona, being rude, not following the instruction)

This project is under an MIT license, which means you may use this dataset in any way you want, as long as it isn't for any illegal and/or unethical purposes. You may use this data, modify it, etc. to fit your criteria. You must credit this github url: johnpaulbin/toki-pona-orpo in your applications.