GoodPrompt: transparent, decentralized, peer-reviewed data submissionA

We introduce an on-chain peer-review process for AI training data. The process enhances AI's transparency, privacy safeguards, accuracy, cost-efficiency, and trustworthiness. The peer-review process is similar to scientific publishing. First, an author submits a data object, which may contain an instructional prompt, context, licenses, and response. Then, reviewers can accept, refer to an expert, ask for improvement, or reject the data object. The final result is a dataset of peer-reviewed data objects transparently stored on a trusted infrastructure such as decentralized ledger. Compensation for authors and reviewers will be governed by a smart-contract based on the authors' and reviewers' reputation and the quality of the submitted objects.

Flow

UI

1. Grant view:

   1. create:
      1. input text: name
      2. input text: description / motivation
      3. input text: constitution / guide
      4. input text x3: examples
      5. button: create a grant
   2. add funds:

2. author view:

   1. manual: enter context,instruction,response,source
   2. LLM parsing: drop in a file (discord or emails) and LLM figures out the list of data points {context,instruction,response,source}
   3. dashboard that shows data, income and progress

3. reviewer view

   1. data object card:
      1. displays the data object {context,instruction,response,source}
      2. button accept (swipe right)
      3. button reject (swipe left)
      4. feedback (expands into a text field for feedback)
      5. refer to an expert button (expands into a field that takes the address or username of the expert, maybe suggestion list)
      6. skip
   2. status info such as rating, stats
   3. dashboard could be a tab or expand from status showing earnings, rating and so on

4. dataset explorer view

   1. search for datasets using graph protocol for a graph by description and tags
   2. show a table of teaching data objects. dispute button next to an object that expands into a feedback field

Bounties

category	sponsor	prize	amount	our fit	dificulty	priority
sign in	Mask Nettwork	RelationService	6000	connect to twitter and build a reputation	6	1
sign in	Cometh	The smoothest UX	3000	easier login and interaction with the chain without signing every transaction	6	3
sign in	Sismo	overall	2500	sign in with min followers on github or twitter and maybe some privacy	10	3
sign in	Polygon ID	privacy id	2500	login	4	3
sign in	Safe	aa	2500	auth-kit login with email, onramp-kit pay using creditcard	5	3
sign in	Safe	protocol	2500	no signing swipes to keep nice ux or sign in the end	4	3
application llm	Compound	Compound community use	5000	create data and grant specifically for compound	4	2
llm & filestorage	Filecoin	Grand Prize	2500	create structured prompt data objects with lillypad fastchat, fvm to programatically create a dataset	10	3
compute	chainlink	functions	1500	use function to run llm or cos similarity	6	5
llm distance	cartesi (no-sponsor)	--	--	check distance of the new submission to encourage diversity and alignment	7	3
notification	Push Protocol	Pushing boundaries / All	1200	notify reviewers that they were chosen through notification or chat. can also use gating for approved reviewers	3	3
notification	Push Protocol	chat	1800	send a message from a reviewer to the author with request for changes	3	3
messages	XMTP	LIT claimable keys	1500	claim a wallet over email. could be used to invite authors and reviewers	6	2
messages	XMTP	best use	250	group chat between authors and reviewers	5	2
voting	UMA	optimistic oracle	5000	submit the data object to UMA so they can handle disputes and voting	10	1
blind voting	aztec	use noir library	2000	use noir library to do blind voting	10	5
manage key-pairs	Lit Protocol	Programmatic Signing	2000	revealing votes after everyone voted or enough time has passed	6	3
data	graph	Best New Graph	1400	index the onchain reviewing data and easily create the training dataset	7	2
data	tableland	studio	1000	put data into sql and createa the training dataset	7	3
UX	Mantle	best UX	1250	deploy on Mantle	2	0
lend	spark	SparkLend	4000	authors and reviewers can get paid in usdc while the grant is in ETH	10	60
cross-chain	connext, hyperlane, wormhole, axelar	--	--	transfer data between chains (fvm), create our own cheaper data chain and transfer from it	10	6

Deployment chains:

chain	prize amount	notes
polygon	2500	public good with account abstraction {safe}
scroll	2000	zk evm should be trivial change of RPC
Mantle	2.500 pool	need to tweet

Tokens:

token	prize amount	notes
ApeCoin	3000	consumer or innovative use case
connext	1500	deploy and use xerc20 (extension of erc20)

.env file in root for deploying smart contract - PRIVATE_KEY= (dataset owner address) ALCHEMY_API_KEY=RHhC7o6uOWm8N0q7zLHJKpwkMZoiKyUG CONTRACT_ADDRESS=0xb3fBCd330499d12C91e30D6e5118e74B355ca135 WALLET_ADDRESS=0x61bbB116879ac169A8D50d20E1E9867Ead619BCC