What would a task/dataset for deception/bad behavoir look like? Maybe some kind of game?
Current theme: default
Less Wrong (text)
Less Wrong (link)
What would a task/dataset for deception/bad behavoir look like? Maybe some kind of game?