-
Notifications
You must be signed in to change notification settings - Fork 26
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Boosting All Rewards Based On Priority #175
Comments
RFC @gentlementlegen I have a feeling that with our current infrastructure it doesn't directly support this plugin to affect the reward amount of others, such as conversation rewards. We need to support outputting rewards from all plugins and aggregating them before sending the sum total to permit generation. ubiquity-os/permit-generation#3 I suppose this proposal implies that we will need to support multiplication as well. |
Assisitive pricing aims the handle pricing labels and allowing users to modify these labels. There is not calculation on the final reward here. I think this should be handled in conversation rewards. |
I see what you mean but logically seems more straightforward if a plugin has the ability to affect all the others' pricing either with addition or multiplication. Then we don't have to change every plugin, we only need to change this one and the permit generation one to support macro change outputs from any plugin I suppose it's simpler to implement if we change every plugin though, just not as elegant. |
Yes I do agree that we should be able to chain these results. It's just that this plugin never had the capability to actually attribute rewards to a user, this was the responsibility of |
If you could write a spec, that would be helpful because I don't have enough context on the problem |
@gentlementlegen can you write the spec? |
@0x4007 Done, let me know if this is clear enough. |
Which plugin handles the final permit generation which combines all the rewards? |
/start |
Tip
|
/stop |
Right now it's Hopefully soon we can upgrade the system to make "permit requests" and every plugin will have a standardized way to send a request to the permit generation module/plugin/endpoint. |
shouldnt this spec be moved to text-conversation rewards for now for a fix and then later a better way can hopefully be found? |
@gentlementlegen rfc |
We can do that, you can also cross-reference issues from another repo, either works. |
I don't understand |
I meant doing the pull-request in |
Like how it's done for devpool directory right? |
Making a pull request to the code base while marking it to resolve the issue in the dedicated issue repo for the devpool directory |
/start |
! ishowvel you were previously unassigned from this task. You cannot be reassigned. |
@ishowvel the deadline is at Tue, Oct 29, 11:04 PM UTC |
is there a way to qa without needing to get an openai api token |
@ishowvel You can either use a free endpoint, use mocks, or hardcode the evaluation for your testing. |
@gentlementlegen would hardcoding the evaluation to be one for each comment be a good idea?
|
also it would very awesome if someone could make the openai implementation optional which will just make all relevances equal to 1 |
https://github.com/ishowvel/text-conversation-rewards/actions/runs/11616689662/job/32350211337 |
@ishowvel You can check the tests where the evaluation is mocked. To test I usually use a free endpoint. You can also start a local mock of OpenAi, you can find a few a GitHub. Not sure for the configuration, you can print the detailed errors to find out. |
We should boost all the rewards proportionally if it's an urgent task. This will incentivize all contributors (not just the assignee) to focus on that task because everybody gets higher rewards for work like reviews.
The higher the priority level the higher the rewards multiplier is for everybody.
example
I should've received a reward closer to $150 in this task, because the priority level was 3.
The text was updated successfully, but these errors were encountered: