SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2202.01771 · 2022

Pre-Trained Language Models for Interactive Decision-Making

arXiv · PDF · Open in the Atlas

Code that ran

13 of 18 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 12 have no contract verdict; 1 failure of our instrument; 5 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2202.01771.json · badge: /api/ran/2202.01771.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
xavierpuigf/virtualhome community (classifier label)
classifier basis not recorded on the link
13 ran · 5 no recorded run
FunctionStatusWhere it lives
build_grid_images Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/demo/generate_snapshot.py
code served (permissive licence) · get_code("adbcc83fbce194cc")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family float, shape (10, 8, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# The function expects 'images' to be a list of lists of numpy arrays
# Each image_step is a list of image arrays that get concatenated along axis 1
# Then all steps are concatenated along axis 0

# Create a simple structure:
# - 2 image steps
# - Each step has 2 camera images
# - Each image is a small numpy array (height, width, channels)

# Using distinct sizes: 4x3 for one dimension, 5x6 for another, etc.
camera_image_1 = np.random.rand(4, 3, 3).astype(np.float32)  # height=4, width=3, channels=3
camera_image_2 = np.random.rand(4, 5, 3).astype(np.float32)  # height=4, width=5, channels=3

image_step_1 = [camera_image_1, camera_image_2]  # Will concatenate along axis 1 (width)

camera_image_3 = np.random.rand(6, 3, 3).astype(np.float32)  # height=6, width=3, channels=3
camera_image_4 = np.random.rand(6, 5, 3).astype(np.float32)  # height=6, width=5, channels=3

image_step_2 = [camera_image_3, camera_image_4]  # Will concatenate along axis 1 (width)

images = [image_step_1, image_step_2]  # Will concatenate along axis 0 (height)

ARGS = {
    "images": images
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family float, shape (10, 8, 3) — shape (10, 8, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
clean_graph Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/demo/utils_demo.py
code served (permissive licence) · get_code("83295b4614f48186")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function expects a graph dict with 'nodes' and 'edges' keys
# It iterates over nodes, removes 'bounding_box' key if present, and copies other keys
# It also copies the edges list

graph = {
    'nodes': [
        {
            'id': 1,
            'label': 'node_1',
            'bounding_box': [0, 0, 10, 10],
            'value': 42
        },
        {
            'id': 2,
            'label': 'node_2',
            'bounding_box': [20, 20, 30, 30],
            'metadata': 'test'
        },
        {
            'id': 3,
            'label': 'node_3',
            'value': 100
        }
    ],
    'edges': [
        (1, 2),
        (2, 3),
        (1, 3)
    ]
}

ARGS = {
    'graph': graph
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
cuboid Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
xavierpuigf/virtualhome/virtualhome/demo/display_graph.py
code served (permissive licence) · get_code("b3ca8a1b900d75f3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Poly3DCollection. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D

# Create a 3D plot
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')

# Define arguments for the cuboid function
# coords: [x, y, z] center coordinates
coords = [0.0, 1.0, 2.0]

# sides: [width, height, depth] dimensions
sides = [2.0, 3.0, 4.0]

ARGS = {
    'ax': ax,
    'coords': coords,
    'sides': sides
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, mpl_toolkits — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Poly3DCollection — a Poly3DCollection output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Poly3DCollection output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_edges_from Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/demo/utils_demo.py
code served (permissive licence) · get_code("44dc4121fc6621b0")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 2, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Construct a minimal graph structure that satisfies the function's requirements
graph = {
    'nodes': [
        {'id': 1, 'label': 'node_1'},
        {'id': 2, 'label': 'node_2'},
        {'id': 3, 'label': 'node_3'},
    ],
    'edges': [
        {'from_id': 1, 'to_id': 2, 'relation_type': 'connects_to'},
        {'from_id': 1, 'to_id': 3, 'relation_type': 'references'},
        {'from_id': 2, 'to_id': 3, 'relation_type': 'depends_on'},
    ]
}

ARGS = {
    'graph': graph,
    'id': 1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
from_hash Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/augment_dataset_exceptions.py
code served (permissive licence) · get_code("058686fb222c2d31")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function takes a precond_tuple parameter and:
# 1. Converts it to a list
# 2. Iterates over elements, where each element is expected to be a tuple/list with:
#    - elem[0]: a key
#    - elem[1]: an iterable of values (some may be tuples)
# 3. Converts any tuple values to lists
# 4. Wraps each {key: values} in a dict
# 5. Returns the list of dicts

# Construct a valid precond_tuple
precond_tuple = (
    ("key1", [1, 2, (3, 4)]),
    ("key2", [5, (6, 7), 8]),
    ("key3", [(9, 10, 11), 12])
)

ARGS = {
    "precond_tuple": precond_tuple
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
getidperobject Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/exception_handler.py
code served (permissive licence) · get_code("05083186903d3268")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Construct valid arguments for getidperobject
object_name = "door handle"
id_env = 5
id_mapping = {
    ("door_handle", 1): 3,
    ("window", 1): 4,
    ("chair", 1): 5
}

ARGS = {
    "object_name": object_name,
    "id_env": id_env,
    "id_mapping": id_mapping
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
insertInstructions Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/augmentation_utils.py
code served (permissive licence) · get_code("f44fbe71aa27eb3d")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function:
# - insert_in: expects iterable of items where each item is indexable (insertval[0] for index, insertval[1] for value)
# - contentold: expects an object with .copy() method and .insert(index, value) method (list-like)
# The function inserts elements from insert_in into content_old at specified positions

# Create insert_in: list of tuples where [0] is position, [1] is value to insert
insert_in = [
    (0, "first"),
    (1, "second"),
    (2, "third")
]

# Create contentold: a simple list that has copy() and insert() methods
contentold = ["a", "b", "c"]

ARGS = {
    "insert_in": insert_in,
    "contentold": contentold
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
obtain_objects_from_message Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/execute_script_utils.py
code served (permissive licence) · get_code("b0f5c4fe4cd3952f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

ARGS = {
    'message': {
        'unplaced': ['object1', 'object2'],
        'missing_destinations': ['dest1'],
        'missing_prefabs': ['prefab1', 'prefab2', 'prefab3']
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
obtain_scene_id_from_path Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/execute_script_utils.py
code served (permissive licence) · get_code("eb99b8eb5c8721af")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

def obtain_scene_id_from_path(path):
    scene_name = [x for x in path.split('/') if 'TrimmedTestScene' in x][0]
    scene_number = int(scene_name.split('TrimmedTestScene')[1].split('_graph')[0])
    return scene_number

# The function takes a path string and extracts a scene number from it
# The path must contain a directory name with 'TrimmedTestScene' followed by a number
path = "/data/TrimmedTestScene42_graph/frame_001.jpg"

ARGS = {
    "path": path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
obtain_script_grounded_in_graph Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/augment_dataset_exceptions.py
code served (permissive licence) · get_code("287b61eceef14fa4")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

# Create stand-in classes for script line objects
class Action:
    def __init__(self, name):
        self.name = name

class ScriptObject:
    def __init__(self, instance):
        self.instance = instance

class ScriptLine:
    def __init__(self, action_name, obj=None, subj=None):
        self.action = Action(action_name)
        self._object = obj
        self._subject = subj
    
    def object(self):
        return self._object
    
    def subject(self):
        return self._subject

# Build the arguments
lines_program = [
    "line 0",
    "line 1",
    "line 2",
    "line 3",
    "line 4",
]

id_mapping = {
    ("ObjectA", "typeA"): "obj_1",
    ("ObjectB", "typeB"): "obj_2",
    ("SubjectX", "typeX"): "subj_1",
}

modified_script = [
    ScriptLine("pick", obj=ScriptObject("obj_1"), subj=ScriptObject("subj_1")),
    ScriptLine("place", obj=ScriptObject("obj_2"), subj=None),
    ScriptLine("move", obj=None, subj=ScriptObject("subj_1")),
]

ARGS = {
    "lines_program": lines_program,
    "id_mapping": id_mapping,
    "modified_script": modified_script,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
parseStrBlock Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/augmentation_utils.py
code served (permissive licence) · get_code("6b36626f01871673")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

# SETUP CODE FOR parseStrBlock

ARGS = {
    "block_str": "[Rinse] <CLEANING SOLUTION> (1)"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
parse_exec_script_file Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/execute_script_utils.py
code served (permissive licence) · get_code("f5247aeac1223674")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import os

# Create a temporary file with the expected format
# The function expects:
# - Line 0: title
# - Line 1: description
# - Line 2: (empty or unused)
# - Line 3: (empty or unused)
# - Line 4+: script with numbered prefixes to remove

temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_file.write("Test Script Title\n")
temp_file.write("This is a test description\n")
temp_file.write("\n")
temp_file.write("\n")
temp_file.write("1. print('hello world')\n")
temp_file.write("2. x = 42\n")
temp_file.write("3. y = x + 8\n")
temp_file.close()

ARGS = {
    'file_name': temp_file.name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
to_hash Ran
ran on a synthesized input
xavierpuigf/virtualhome/virtualhome/dataset_utils/augment_dataset_exceptions.py
code served (permissive licence) · get_code("998d6b9491c17a70")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

# Analyzing the function to_hash:
# - Takes precond_list as parameter
# - Copies it, iterates through elements
# - Each element is treated as a dict (list(elem)[0] gets first key)
# - Gets values from that key
# - Converts any list values to tuples
# - Builds tuple_dict as (key, tuple_of_values)
# - Returns sorted tuple of all tuple_dicts

# Create a valid precond_list with dict elements
# Each element should be a dict with one key mapping to a list/tuple of values

precond_list = [
    {"param_a": [1, 2, 3]},
    {"param_b": [[4, 5], 6, 7]},
    {"param_c": [8, [9, 10]]}
]

ARGS = {
    "precond_list": precond_list
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_nodes Not yet run
unverified (harvested, no recorded run)
xavierpuigf/virtualhome/virtualhome/demo/utils_demo.py
code served (permissive licence) · get_code("8e3413a872edc331")
How it ran

no invocation recorded.

get_preconds_script Not yet run
unverified (harvested, no recorded run)
xavierpuigf/virtualhome/virtualhome/dataset_utils/add_preconds.py
code served (permissive licence) · get_code("1aef779a0ec40326")
How it ran

no invocation recorded.

get_xyz_mouse_click Not yet run
unverified (harvested, no recorded run)
xavierpuigf/virtualhome/virtualhome/demo/display_graph.py
code served (permissive licence) · get_code("fc264d88eed3353e")
How it ran

no invocation recorded.

hasProperty Not yet run
unverified (harvested, no recorded run)
xavierpuigf/virtualhome/virtualhome/dataset_utils/augmentation_utils.py
code served (permissive licence) · get_code("ff70381f7e6ddae8")
How it ran

no invocation recorded.

parseException Not yet run
unverified (harvested, no recorded run)
xavierpuigf/virtualhome/virtualhome/dataset_utils/exception_handler.py
code served (permissive licence) · get_code("b58216b38e8b12f8")
How it ran

no invocation recorded.

Also present in other papers' code

Of 18 samples lifted from this paper's linked repositories, 18 carry a function identity and 0 do not. Of the 18 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 0 later) and 15 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
cuboid 1 other paper
1 dated earlier
2104.02215 (2021-04)
When Pigs Fly: Contextual Reasoning in Synthetic and Natural Scenes
find_nodes 1 other paper
1 dated earlier
2104.02215 (2021-04)
When Pigs Fly: Contextual Reasoning in Synthetic and Natural Scenes
get_xyz_mouse_click 1 other paper
1 dated earlier
2104.02215 (2021-04)
When Pigs Fly: Contextual Reasoning in Synthetic and Natural Scenes

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Language model (LM) pre-training is useful in many language processing tasks. But can pre-trained LMs be further leveraged for more general machine learning problems? We propose an approach for using LMs to scaffold learning and generalization in general sequential decision-making problems. In this approach, goals and observations are represented as a sequence of embeddings, and a policy network initialized with a pre-trained LM predicts the next action. We demonstrate that this framework enables effective combinatorial generalization across different environments and supervisory modalities. We begin by assuming access to a set of expert demonstrations, and show that initializing policies with LMs and fine-tuning them via behavior cloning improves task completion rates by 43.6% in the VirtualHome environment. Next, we integrate an active data gathering procedure in which agents iteratively interact with the environment, relabel past "failed" experiences with new goals, and update their policies in a self-supervised loop. Active data gathering further improves combinatorial generalization, outperforming the best baseline by 25.1%. Finally, we explain these results by investigating three possible factors underlying the effectiveness of the LM-based policy. We find that sequential input representations (vs. fixed-dimensional feature vectors) and LM-based weight initialization are both important for generalization. Surprisingly, however, the format of the policy inputs encoding (e.g. as a natural language string vs. an arbitrary sequential encoding) has little influence. Together, these results suggest that language modeling induces representations that are useful for modeling not just language, but also goals and plans; these representations can aid learning and generalization even outside of language processing.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2202.01771")
get_code_for_paper("2202.01771")
have("2202.01771")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2202.01771.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2202.01771.svg)](https://syntology.ai/paper/2202.01771)

Connect an agent — have() is free.