To start a dialog type or choose a /begin command .
You will be connected to a peer or, if no peer is available at the moment, you’ll receive the message from @ConvaiBotPlease wait for you peer.
Peer might be a bot or another human evaluator.
Warning! Since your peer is either human, or bot she/it could use words or language that considered profane, vulgar, or offensive by some readers. The organizers of ConvAI are not responsible for dialogs contents. It you have spotted a serious issue, please contact organizers about it by email firstname.lastname@example.org.
After you were connected with your peer you will receive a starting message - a passage or two from a Wikipedia article.
Your task is to discuss the content of a presented passage with the peer and score her/his replies.
Please score every utterance of your peer with a ‘thumb UP’ button if you like it, and ‘thumb DOWN’ button in the opposite case.
To finish the conversation type or choose a command /end.
When the conversation is finished, you will receive a request from @ConvaiBot to score the overall quality of the dialog along three dimensions:
quality - how much are you satisfied with the whole conversation?
breadth - in your opinion was a topic discussed thoroughly or just from one side?
engagement - was it interesting to participate in this conversation?
If your peer ends the dialog before you, you will also receive a scoring request from @ConvaiBot.
Your conversations with a peer will be recorded for further use. By starting a chat you give permission to record your conversation and to release your anonymised conversation data publicly under Apache License Version 2.0.