SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1806.00608 · 2018

GamePad: A Learning Environment for Theorem Proving

arXiv · PDF · Open in the Atlas

Code that ran

9 of 11 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 9 have no contract verdict; 0 failures of our instrument; 2 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/1806.00608.json · badge: /api/ran/1806.00608.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
ml4tp/gamepad canonical (classifier label)
classifier basis: census
9 ran · 2 no recorded run
FunctionStatusWhere it lives
get_other Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/ml/fold_model.py
code served (permissive licence) · get_code("5c8e42c87c25fb8c")
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)

# Analyze the function:
# - l: a list (used with len(), indexing [], and slicing)
# - index: an integer (used with % operator)

# Create a simple list with distinct elements to verify behavior
l = [10, 20, 30, 40, 50]
index = 2

ARGS = {
    "l": l,
    "index": index
}
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
load_tactr_stats Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/exp_tactr_stats.py
code served (permissive licence) · get_code("ebe42ade6a0004d2")
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, 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 json
import tempfile
import os

# Create a temporary file with the expected format
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_filename = temp_file.name

# Write sample data matching the function's expected format
temp_file.write("LEMMA INFO\n")
temp_file.write("UNIQUE-SORT 5\n")
temp_file.write("UNIQUE-CONST 10\n")
temp_file.write("UNIQUE-IND 3\n")
temp_file.write("UNIQUE-CONID 7\n")
temp_file.write("UNIQUE-EVAR 2\n")
temp_file.write("UNIQUE-FIX 1\n")
temp_file.write("NUM_IARG\n")
temp_file.write("NUM_ARGS\n")
temp_file.write('{"lemma": "test_lemma_1", "info": {"data": "value1"}}\n')
temp_file.write('{"lemma": "test_lemma_2", "info": {"data": "value2"}}\n')
temp_file.write("TOTAL\n")
temp_file.close()

ARGS = {
    'filename': temp_filename
}
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, ndim 1, dtype family float — a tuple 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
merge_hist Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/lib/myutil.py
code served (permissive licence) · get_code("f851c0e141f110e6")
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)

# merge_hist takes two dictionaries and merges them
# Looking at the function body:
# - It iterates over items in dict1 and dict2
# - For matching keys, it adds values together (v += v means numeric addition)
# - For new keys, it assigns the value
# This suggests the dictionaries contain numeric values (could be ints, floats, or tensors)

dict1 = {
    'loss': 0.5,
    'accuracy': 0.95,
    'steps': 100
}

dict2 = {
    'loss': 0.3,
    'validation_accuracy': 0.92,
    'steps': 50
}

ARGS = {
    'dict1': dict1,
    'dict2': dict2
}
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
merge_hists Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/lib/myutil.py
code served (permissive licence) · get_code("7d7a82026c846b65")
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
import numpy as np

torch.manual_seed(0)

# Analyze the function: merge_hists takes a list of dicts
# Each dict should have string keys and numeric values (since they're added with +=)
# The function merges multiple dicts by combining values for matching keys

dict1 = {
    "loss": 0.5,
    "accuracy": 0.9,
    "iterations": 100
}

dict2 = {
    "loss": 0.3,
    "f1_score": 0.85,
    "iterations": 50
}

dict3 = {
    "accuracy": 0.1,
    "iterations": 25,
    "precision": 0.92
}

dicts = [dict1, dict2, dict3]

ARGS = {
    "dicts": dicts
}
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 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
pp_tab Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/lib/myutil.py
code served (permissive licence) · get_code("29dafeed16d661b7")
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 str. 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

def pp_tab(tab, s):
    return tab * ' ' + s

ARGS = {
    'tab': 3,
    's': 'hello'
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
return_as Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/lib/sexpdata.py
code served (permissive licence) · get_code("6d0d15d341f47b06")
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 function. 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 functools
import torch

def return_as(converter):
    """
    Decorator to convert result of a function.
    It is just a function composition. The following two codes are
    equivalent.
    Using `@return_as`::
        @return_as(converter)
        def generator(args):
            ...
        result = generator(args)
    Manually do the same::
        def generator(args):
            ...
        result = converter(generator(args))
    Example:
    >>> @return_as(list)
    ... def f():
    ...     for i in range(3):
    ...         yield i
    ...
    >>> f()  # this gives a list, not an iterator
    [0, 1, 2]
    """

    def wrapper(generator):

        @functools.wraps(generator)
        def func(*args, **kwds):
            return converter(generator(*args, **kwds))
        return func
    return wrapper


# Setup code
def simple_generator():
    """A simple generator function to decorate."""
    for i in range(3):
        yield i


ARGS = {
    "converter": list
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
seq_embed Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/ml/fold_model.py
code served (permissive licence) · get_code("15f3646408acd1cf")
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 torch.nn as nn

torch.manual_seed(0)

class MockFolder:
    def __init__(self):
        self.operations = []
    
    def add(self, op_name, *args):
        self.operations.append((op_name, args))
        # Return a dummy tensor to simulate layer output
        return torch.randn(2, 4)

class MockCell(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(4, 4)
    
    def forward(self, x, hidden):
        return self.linear(x + hidden)

class MockInputDropout(nn.Module):
    def __init__(self):
        super().__init__()
    
    def forward(self, x):
        return x

class MockLayerNorm(nn.Module):
    def __init__(self):
        super().__init__()
        self.ln = nn.LayerNorm(4)
    
    def forward(self, x):
        return self.ln(x)

class MockIdentity(nn.Module):
    def __init__(self):
        super().__init__()
    
    def forward(self, x):
        return x

ARGS = {
    "name": "gru",
    "folder": MockFolder(),
    "xs": [torch.randn(2, 4), torch.randn(2, 4), torch.randn(2, 4)],
    "init": torch.randn(2, 4),
    "get_hiddens": True,
    "ln": True,
    "input_dropout": True,
    "conclu_pos": 0
}
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 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
seq_sigmoid_attn_embed Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/ml/fold_model.py
code served (permissive licence) · get_code("41530535e0588bd4")
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 Tensor, ndim 2, dtype family float, shape (2, 4) 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
import torch.nn as nn

torch.manual_seed(0)

class MockFolder:
    """Mimics a folder object that has an add method"""
    def add(self, operation_name, *args):
        # Returns a transformed tensor based on the operation
        if operation_name == 'input_dropout_f':
            # Takes one tensor, returns a tensor
            return args[0]
        elif operation_name == 'attn_q_f':
            # Takes one tensor (conclu), returns a tensor
            return args[0]
        elif operation_name == 'attn_identity':
            # Takes sv_init, returns it as-is
            return args[0]
        elif operation_name == 'attn_sv_f':
            # Takes q, x, sv - returns a tensor
            q, x, sv = args
            return sv + x  # Simple combination
        else:
            return args[0]

# Create small tensors with distinct dimensions
batch_size = 2
dim = 4

folder = MockFolder()
xs = [
    torch.randn(batch_size, dim),
    torch.randn(batch_size, dim),
    torch.randn(batch_size, dim)
]
sv_init = torch.randn(batch_size, dim)
ln = None  # Not used in the function body
input_dropout = False  # Can be True or False, function handles both
conclu_pos = 1  # Index into xs list

ARGS = {
    'folder': folder,
    'xs': xs,
    'sv_init': sv_init,
    'ln': ln,
    'input_dropout': input_dropout,
    'conclu_pos': conclu_pos
}
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 Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) 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
sexpr_unpack Ran
ran on a synthesized input
ml4tp/gamepad/gamepad/lib/mysexpr.py
code served (permissive licence) · get_code("1733bf935dfc369a")
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

class MockVal:
    def __init__(self, value):
        self._val = value

# Test case 1: sexpr is a sequence with a tag and body
tag_obj = MockVal("tag1")
body_items = [MockVal("item1"), MockVal("item2")]
sexpr_list = [tag_obj] + body_items

ARGS = {
    "sexpr": sexpr_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
tounicode Not yet run
unverified (harvested, no recorded run)
ml4tp/gamepad/gamepad/lib/sexpdata.py
code served (permissive licence) · get_code("11fea3b1e5d3e940")
How it ran

no invocation recorded.

uformat Not yet run
unverified (harvested, no recorded run)
ml4tp/gamepad/gamepad/lib/sexpdata.py
code served (permissive licence) · get_code("54423a2912e662b6")
How it ran

no invocation recorded.

Also present in other papers' code

Of 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 11 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.

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

In this paper, we introduce a system called GamePad that can be used to explore the application of machine learning methods to theorem proving in the Coq proof assistant. Interactive theorem provers such as Coq enable users to construct machine-checkable proofs in a step-by-step manner. Hence, they provide an opportunity to explore theorem proving with human supervision. We use GamePad to synthesize proofs for a simple algebraic rewrite problem and train baseline models for a formalization of the Feit-Thompson theorem. We address position evaluation (i.e., predict the number of proof steps left) and tactic prediction (i.e., predict the next proof step) tasks, which arise naturally in tactic-based theorem proving.

For agents

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

get_harvested_code_for_paper("1806.00608")
get_code_for_paper("1806.00608")
have("1806.00608")

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

curl https://syntology.ai/api/ran/1806.00608.json

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

[![Syntology run record](https://syntology.ai/api/ran/1806.00608.svg)](https://syntology.ai/paper/1806.00608)

Connect an agent — have() is free.