SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2006.09265 · ICLR · 2021

ISARSTEP: A BENCHMARK FOR HIGH-LEVEL MATHEMATICAL REASONING

Wenda Li, Lei Yu, Yuhuai Wu, Lawrence Paulson

arXiv · PDF · Open in the Atlas

Code that ran

2 of 8 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 2 have no contract verdict; 0 failures of our instrument; 6 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/2006.09265.json · badge: /api/ran/2006.09265.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
reactive-systems/ml2 community
Papers with Code archive
2 ran · 5 no recorded run
reactive-systems/circuit-repair community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
BeamSearch Ran
ran on a synthesized input · constructed an object; no result was computed
reactive-systems/ml2/ml2/models/tf_hierarchical_transformer.py
code served (permissive licence) · get_code("f3ad3707e2c07557")
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 an instance of its own class, BeamSearch: the run constructed the object and did not call it, so no result was computed. 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
tf.random.set_seed(0)

# Define a simple logits function that the BeamSearch will call
def logits_fn(ids, cache):
    """
    Mock logits function that returns logits for vocabulary predictions.
    Args:
        ids: token ids of shape (batch_size * beam_size, seq_len)
        cache: dictionary of cached tensors
    Returns:
        logits: tensor of shape (batch_size * beam_size, vocab_size)
        new_cache: updated cache dictionary
    """
    batch_beam_size = tf.shape(ids)[0]
    vocab_size = 16
    logits = tf.random.normal([batch_beam_size, vocab_size], dtype=tf.float32)
    return logits, cache

# Create parameters dictionary
params = {
    'alpha': 0.6,
    'beam_size': 3,
    'dtype_float': tf.float32,
    'target_eos_id': 2,
    'max_decode_length': 10,
    'target_vocab_size': 16
}

# Create initial_ids: shape (batch_size,) with token ids
batch_size = 2
initial_ids = tf.constant([1, 1], dtype=tf.int32)

# Create initial_cache: dictionary of cached tensors
# This will be expanded by expand_to_beam_size in get_initial_state
initial_cache = {
    'encoder_output': tf.random.normal([batch_size, 5, 8], dtype=tf.float32),
    'encoder_state': tf.random.normal([batch_size, 6], dtype=tf.float32)
}

# Construct ARGS dictionary
ARGS = {
    'logits_fn': logits_fn,
    'batch_size': batch_size,
    'params': params
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type BeamSearch — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a BeamSearch 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
architecture_is_apple_arm Ran
ran on a synthesized input
reactive-systems/ml2/ml2/models/tf_hierarchical_transformer.py
code served (permissive licence) · get_code("7ac4bc1ee866b584")
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 bool, ndim 0, 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 platform

# This function takes no parameters, so ARGS should be an empty dict
ARGS = {}
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 bool, ndim 0, dtype family float — no output shape is stated: observe_signature.py could not observe it (no_inputs)
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
HierarchicalTransformer Not yet run
unverified (harvested, no recorded run)
reactive-systems/ml2/ml2/models/tf_hierarchical_transformer.py
code served (permissive licence) · get_code("ada4f838ec590007")
How it ran

no invocation recorded.

HierarchicalTransformer Not yet run
unverified (harvested, no recorded run)
reactive-systems/circuit-repair/ml2/models/hierarchical_transformer_2d.py
code served (permissive licence) · get_code("455e8cd540f0a9e4")
How it ran

no invocation recorded.

TransformerDecoder Not yet run
unverified (harvested, no recorded run)
reactive-systems/ml2/ml2/models/tf_hierarchical_transformer.py
code served (permissive licence) · get_code("bb5bd0a774421a8c")
How it ran

no invocation recorded.

TransformerDecoderLayer Not yet run
unverified (harvested, no recorded run)
reactive-systems/ml2/ml2/models/tf_hierarchical_transformer.py
code served (permissive licence) · get_code("5026c83d97c9ffb8")
How it ran

no invocation recorded.

TransformerEncoder Not yet run
unverified (harvested, no recorded run)
reactive-systems/ml2/ml2/models/tf_hierarchical_transformer.py
code served (permissive licence) · get_code("2d6d9a5e8486aa8e")
How it ran

no invocation recorded.

TransformerEncoderLayer Not yet run
unverified (harvested, no recorded run)
reactive-systems/ml2/ml2/models/tf_hierarchical_transformer.py
code served (permissive licence) · get_code("96e9ed72cac68cc4")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 0 later) and 2 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
BeamSearch
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
TransformerDecoder
brick of this paper's assembly: model
1 other paper
1 dated earlier · 1 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
TransformerDecoderLayer
brick of this paper's assembly: model
1 other paper
1 dated earlier · 1 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
TransformerEncoder
brick of this paper's assembly: model
1 other paper
1 dated earlier · 1 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
TransformerEncoderLayer
brick of this paper's assembly: model
1 other paper
1 dated earlier · 1 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
architecture_is_apple_arm
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need

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

A well-defined benchmark is essential for measuring and accelerating research progress of machine learning models. In this paper, we present a benchmark for high-level mathematical reasoning and study the reasoning capabilities of neural sequence-to-sequence models. We build a non-synthetic dataset from the largest repository of proofs written by human experts in a theorem prover. The dataset has a broad coverage of undergraduate and research-level mathematical and computer science theorems. In our defined task, a model is required to fill in a missing intermediate proposition given surrounding proofs. This task provides a starting point for the long-term goal of having machines generate human-readable proofs automatically. Our experiments and analysis reveal that while the task is challenging, neural models can capture non-trivial mathematical reasoning. We further design a hierarchical transformer that outperforms the transformer baseline. The dataset and models are available from: https://github.com/ Wenda302/IsarStep.

For agents

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

get_harvested_code_for_paper("2006.09265")
get_code_for_paper("2006.09265")
have("2006.09265")

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

curl https://syntology.ai/api/ran/2006.09265.json

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

[![Syntology run record](https://syntology.ai/api/ran/2006.09265.svg)](https://syntology.ai/paper/2006.09265)

Connect an agent — have() is free.