Text this: From T-mazes to labyrinths: Learning from model-based feedback

 ______    _    _               _____    __   __  
|      \\ | || | ||   ____     |  ___||  \ \\/ // 
|  --  // | || | ||  |    \\   | ||__     \   //  
|  --  \\ | \\_/ ||  | [] ||   | ||__     / . \\  
|______//  \____//   |  __//   |_____||  /_//\_\\ 
`------`    `---`    |_|`-`    `-----`   `-`  --` 
                     `-`