Hey all,
I need some help, if you guys could fill out this form and copy/paste or transcribe gold spam messages you see (as exactly as possible) in shout/normal chat this would help me tremendously with a little pet project I’d like to try out.
Here’s the survey, so please fill it out if you can!
@moderators if imc could help out with this effort (as I’m sure you guys are sitting on a ton of data), that would also help tremendously.
The idea is to try a neural network classifier which should be more robust than naive bayes style classifiers. I probably won’t open source the solution sadly since showing it means spammers can reverse engineer how to get messages across.




