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

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