RewardPrediction / data /textworld /194 /positive.json
YijunShen's picture
Initial commit
b087771 verified
{
"goal_description": "retrieve key from locker in scullery. unlock safe in scullery. open safe. retrieve worm from safe. place worm on floor of scullery.",
"trajectory": [
{
"action": "take key from locker",
"observation": "You take the key from the locker.",
"reward": {
"raw": 0.0,
"shaped": 0.2,
"is_expert": true
}
},
{
"action": "unlock safe with key",
"observation": "You unlock the safe.",
"reward": {
"raw": 0.0,
"shaped": 0.4,
"is_expert": true
}
},
{
"action": "open safe",
"observation": "You open the safe, revealing a worm.",
"reward": {
"raw": 0.0,
"shaped": 0.6,
"is_expert": true
}
},
{
"action": "take worm from safe",
"observation": "You take the worm from the safe.",
"reward": {
"raw": 0.0,
"shaped": 0.8,
"is_expert": true
}
},
{
"action": "drop worm",
"observation": "You drop the worm on the ground. Your score has just gone up by one point. *** The End *** You scored 1 out of a possible 1, in 6 turns. Would you like to RESTART, RESTORE a saved game, QUIT or UNDO the last command?",
"reward": {
"raw": 1.0,
"shaped": 1.0,
"is_expert": true
}
}
]
}