Credit Score
The credit score in this system is determined based on a combination of factors related to the user's API usage. The calculation involves several steps, each contributing to the final score:
1. Base Score:
The credit score starts with a base value of 600 points if the user has engaged with the system (i.e., purchased at least 1 USDT in API tokens). If the user has never interacted with the system, the base score will be set to 400.
2. Model Usage:
WitchBot Usage: If the user has chatted with WitchBot, user get some credit addition if questions answered correctly, user can get up to 25 points.
Pilot Kombat Usage: If the user has used Pilot Kombat, each 1000 points lead to an additional 1 credit, user can get up to 25 points.
Advanced Model Usage (Context Pilot VSCode Plugin):
If the user has collectively used advanced models more than 10 times, an additional 25 points are added.
If the usage exceeds 100 times, another 25 points are added.
3. Usage Consistency:
The score considers how consistently the user interacts with the API over time. If the usage is spread evenly across different days, the credit score can be increased by up to 50 points. This is determined by analyzing the standard deviation of daily API usage, with more consistent usage leading to a higher score.
4. Twitter Intent Adjustment:
The adjustment to the credit score based on Twitter intent is determined by analyzing the user's interaction metrics, if user clicked any of like, tweet, retweet
if user is honest, and really took the action at tweet, they get up to 10 more credits
if user is dis-honest, and really took the action at tweet, they get up to 10 minus credits
5. Adjustments for Model Score:
We will assign a score to the model usage in the VSCode plugin, based on the average score, the credit score will be adjusted
if the model usage is professional user could get up to 10 bonus points
if the model usage is more of a replay attack like or meaningless, there would be up to 5 points punishment
6. Credit Check (TODO):
More credit score will be added or deducted based on the user's borrow and payback history. This feature is planned for future implementation.
Final Credit Score:
The final credit score is the sum of the base score plus any additional points earned through model usage, Twitter intent adjustments, and consistency. The score provides an overall measure of the user's reliability and engagement with the API.
Last updated