Reinforcement World Model Learning for LLM-based Agents
Paper • 2602.05842 • Published • 28
data_source stringclasses 1
value | prompt listlengths 2 2 | reward_model dict | extra_info dict |
|---|---|---|---|
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 1. On the cabinet 1, you see nothing."
} | {
"record_id": "task_0000/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at fridge 1. The fridge 1 is closed."
} | {
"record_id": "task_0002/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a butterknife 1, a dishsponge 1, and a mug 1."
} | {
"record_id": "task_0003/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at sidetable 1. On the sidetable 1, you see a houseplant 1."
} | {
"record_id": "task_0006/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the fridge 1. The fridge 1 is open. In it, you see a apple 3, a bowl 3, a bread 1, and a potato 1."
} | {
"record_id": "task_0002/rollout_00/step_01"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the cabinet 1. The cabinet 1 is open. In it, you see nothing."
} | {
"record_id": "task_0004/rollout_00/step_01"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at sidetable 2. On the sidetable 2, you see a keychain 3, and a laptop 1."
} | {
"record_id": "task_0006/rollout_00/step_01"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You close the fridge 1."
} | {
"record_id": "task_0002/rollout_00/step_02"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at sidetable 3. On the sidetable 3, you see a book 4."
} | {
"record_id": "task_0006/rollout_00/step_02"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a apple 2, a cup 1, a egg 3, a knife 1, a mug 1, and a saltshaker 1."
} | {
"record_id": "task_0002/rollout_00/step_03"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a candle 3, a cloth 1, a spraybottle 2, a tissuebox 3, and a tissuebox 1."
} | {
"record_id": "task_0007/rollout_00/step_03"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You turn on the desklamp 1."
} | {
"record_id": "task_0001/rollout_00/step_04"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 3. The cabinet 3 is closed."
} | {
"record_id": "task_0004/rollout_00/step_04"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at armchair 1. On the armchair 1, you see a remotecontrol 1."
} | {
"record_id": "task_0006/rollout_00/step_04"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the cabinet 3. The cabinet 3 is open. In it, you see nothing."
} | {
"record_id": "task_0004/rollout_00/step_05"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You cool the plate 1 using the fridge 1."
} | {
"record_id": "task_0005/rollout_00/step_05"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 2. On the countertop 2, you see a lettuce 1, a soapbottle 2, and a tomato 4."
} | {
"record_id": "task_0003/rollout_00/step_07"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You close the drawer 1."
} | {
"record_id": "task_0007/rollout_00/step_07"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a cloth 1, a spraybottle 2, a tissuebox 3, and a tissuebox 1."
} | {
"record_id": "task_0007/rollout_00/step_08"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "There's nothing special about tomato 4."
} | {
"record_id": "task_0003/rollout_00/step_11"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the cabinet 6. The cabinet 6 is open. In it, you see a plate 2."
} | {
"record_id": "task_0004/rollout_00/step_12"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 7. The cabinet 7 is closed."
} | {
"record_id": "task_0004/rollout_00/step_14"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at drawer 4. On the drawer 4, you see nothing."
} | {
"record_id": "task_0007/rollout_00/step_14"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the cabinet 2. The cabinet 2 is open. In it, you see a saltshaker 2."
} | {
"record_id": "task_0002/rollout_00/step_16"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 3. On the countertop 3, you see a butterknife 1, a fork 1, a knife 2, a lettuce 2, a spatula 3, and a tomato 1."
} | {
"record_id": "task_0002/rollout_00/step_17"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the drawer 5. The drawer 5 is open. In it, you see a tissuebox 2."
} | {
"record_id": "task_0007/rollout_00/step_17"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You are carrying: a tomato 4."
} | {
"record_id": "task_0003/rollout_00/step_18"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "There's nothing special about tomato 1."
} | {
"record_id": "task_0002/rollout_00/step_19"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the cabinet 5. The cabinet 5 is open. In it, you see a winebottle 1."
} | {
"record_id": "task_0000/rollout_00/step_20"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the fridge 1. The fridge 1 is open. In it, you see a apple 1, a bread 1, a mug 2, a potato 1, a tomato 3, and a tomato 2."
} | {
"record_id": "task_0003/rollout_00/step_20"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 8. The cabinet 8 is closed."
} | {
"record_id": "task_0004/rollout_00/step_20"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at drawer 7. The drawer 7 is closed."
} | {
"record_id": "task_0007/rollout_00/step_20"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "The cabinet 8 is closed."
} | {
"record_id": "task_0004/rollout_00/step_21"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "There's nothing special about tissuebox 2."
} | {
"record_id": "task_0007/rollout_00/step_21"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at drawer 8. The drawer 8 is closed."
} | {
"record_id": "task_0007/rollout_00/step_22"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a bowl 1, a bread 1, a cup 1, a glassbottle 1, and a tomato 1."
} | {
"record_id": "task_0000/rollout_00/step_24"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "This is a hot tomato 1."
} | {
"record_id": "task_0002/rollout_00/step_24"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You are not carrying anything."
} | {
"record_id": "task_0003/rollout_00/step_28"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "On the dresser 1, you see a alarmclock 1, a bowl 1, a cd 1, a cellphone 2, a laptop 2, and a pen 1."
} | {
"record_id": "task_0001/rollout_00/step_29"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at coffeemachine 1. On the coffeemachine 1, you see a mug 1."
} | {
"record_id": "task_0010/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a bread 2, a cup 1, a dishsponge 1, a egg 2, a fork 1, a knife 1, a lettuce 1, a plate 1, a potato 1, a saltshaker 2, a spatula 1, and a spoon 2."
} | {
"record_id": "task_0013/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at fridge 1. The fridge 1 is closed."
} | {
"record_id": "task_0014/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 1. On the cabinet 1, you see nothing."
} | {
"record_id": "task_0015/rollout_00/step_00"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You are facing the toilet 1. Next to it, you see nothing."
} | {
"record_id": "task_0009/rollout_00/step_01"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the drawer 1. The drawer 1 is open. In it, you see a candle 1."
} | {
"record_id": "task_0012/rollout_00/step_01"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the cabinet 1. The cabinet 1 is open. In it, you see a saltshaker 1."
} | {
"record_id": "task_0011/rollout_00/step_02"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a candle 3, a dishsponge 1, a soapbar 1, and a toiletpaper 2."
} | {
"record_id": "task_0012/rollout_00/step_02"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a bread 1, a butterknife 1, a fork 3, a fork 2, a ladle 2, a ladle 1, a lettuce 1, a pan 1, a spoon 3, and a spoon 1."
} | {
"record_id": "task_0014/rollout_00/step_04"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "The cabinet 2 is closed."
} | {
"record_id": "task_0009/rollout_00/step_05"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You open the fridge 1. The fridge 1 is open. In it, you see a bread 3, a lettuce 2, and a lettuce 1."
} | {
"record_id": "task_0010/rollout_00/step_05"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "Available commands:\n look: look around your current location\n inventory: check your current inventory\n go to (receptacle): move to a receptacle\n open (receptacle): open a receptacle\n close (... | {
"record_id": "task_0012/rollout_00/step_05"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You cool the mug 1 using the fridge 1."
} | {
"record_id": "task_0010/rollout_00/step_06"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You are carrying: a candle 3."
} | {
"record_id": "task_0012/rollout_00/step_07"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at bathtubbasin 1. On the bathtubbasin 1, you see nothing."
} | {
"record_id": "task_0008/rollout_00/step_08"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "There's nothing special about candle 3."
} | {
"record_id": "task_0012/rollout_00/step_08"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 1. On the cabinet 1, you see a dishsponge 1."
} | {
"record_id": "task_0014/rollout_00/step_08"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "The fridge 1 is closed."
} | {
"record_id": "task_0010/rollout_00/step_09"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "The fridge 1 is closed."
} | {
"record_id": "task_0010/rollout_00/step_11"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 2. On the cabinet 2, you see a bowl 1."
} | {
"record_id": "task_0013/rollout_00/step_11"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "The drawer 2 is closed."
} | {
"record_id": "task_0008/rollout_00/step_13"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at cabinet 4. On the cabinet 4, you see a peppershaker 2."
} | {
"record_id": "task_0011/rollout_00/step_13"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "You arrive at countertop 1. On the countertop 1, you see a candle 1, a soapbottle 1, a spraybottle 3, and a spraybottle 1."
} | {
"record_id": "task_0009/rollout_00/step_14"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "On the countertop 1, you see a candle 3, a dishsponge 1, a soapbar 1, and a toiletpaper 2."
} | {
"record_id": "task_0012/rollout_00/step_15"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "The fridge 1 is closed."
} | {
"record_id": "task_0010/rollout_00/step_16"
} |
alfworld_rwml | [
{
"role": "system",
"content": "You are an expert agent operating in the ALFRED Embodied\nEnvironment. Given the task interaction history and a potential action, predict the immediate next\nobservation after taking that action. You should first briefly reason step-by-step about the\nprevious steps and curre... | {
"style": "rule",
"ground_truth": "The fridge 1 is closed."
} | {
"record_id": "task_0010/rollout_00/step_19"
} |
End of preview. Expand in Data Studio
E2L RWML ALFWorld data
RWML (arXiv:2602.05842) reproduction data from
Ch1nyzzz/e2-learning, collected with
Qwen/Qwen2.5-7B-Instruct rollouts on ALFWorld:
rwml_grpo/train.parquet,rwml_grpo/val.parquet— prompts for RWML GRPO training (filtered merged-10k corpus);rwml_tau_calibration.json— calibrated embedding-distance threshold tau_d;rwml_alfworld_qwen25_7b_validation_merged10k.jsonl— validation transitions.
Paths mirror the source repository's data/ layout; pull them back with
scripts/pull_hf_artifacts.sh.
- Downloads last month
- 32