SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1803.02893 · 2018

An efficient framework for learning sentence representations

arXiv · PDF · Open in the Atlas

Code that ran

12 of 21 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 11 have no contract verdict; 1 failure of our instrument; 9 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/1803.02893.json · badge: /api/ran/1803.02893.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
google/embedding-tests community (classifier label)
classifier basis not recorded on the link
6 ran · 5 no recorded run
asappresearch/dialog-intent-induction community (classifier label)
classifier basis not recorded on the link
6 ran · 3 no recorded run
coleridge72/quick_thoughts reimplementation (classifier label)
classifier basis: classifier_2025
0 ran · 1 no recorded run
FunctionStatusWhere it lives
build_ae_model Ran
ran on a synthesized input
google/embedding-tests/attribute/models.py
code served (permissive licence) · get_code("aa39e8c335b8b839")
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 tensorflow as tf

ARGS = {
    'num_attr': 5,
    'hidden_size': 8,
    'input_size': 4,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — 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
build_model Ran
ran on a synthesized input
google/embedding-tests/attribute/models.py
code served (permissive licence) · get_code("b4689ebdf48c8bdf")
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 tensorflow as tf

ARGS = {
    'num_attr': 5,
    'hidden_size': 0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensorflow — 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
build_vocabulary Ran
ran on a synthesized input
google/embedding-tests/authorship_text.py
code served (permissive licence) · get_code("1699bd3e697065eb")
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.

def build_vocabulary(train_sents, test_sents, unlabeled_sents=(), min_word_count=2):
    all_train_words = [w for sent in train_sents for w in sent]
    all_test_words = [w for sent in test_sents for w in sent]
    all_words = all_train_words + all_test_words
    if len(unlabeled_sents):
        all_words += [w for sent in unlabeled_sents for w in sent]
    word_count = Counter(all_words)
    vocab = dict()
    idx = 1
    for word, count in word_count.most_common():
        if count >= min_word_count:
            vocab[word] = idx
            idx += 1
    return vocab

train_sents = [
    ["hello", "world", "hello"],
    ["foo", "bar", "foo"],
]

test_sents = [
    ["hello", "baz"],
    ["world", "qux"],
]

unlabeled_sents = [
    ["hello", "hello", "hello"],
    ["rare"],
]

min_word_count = 2

ARGS = {
    "train_sents": train_sents,
    "test_sents": test_sents,
    "unlabeled_sents": unlabeled_sents,
    "min_word_count": min_word_count,
}
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 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
create_lookup_tables Ran
ran; our contract draft was wrong, not the code
google/embedding-tests/train_word_embedding_dp.py
code served (permissive licence) · get_code("f9c19e8923763bb4")
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.

from collections import Counter

# Construct a simple list of words for the function to process
words = ['hello', 'world', 'hello', 'python', 'world', 'hello', 'code', 'python', 'python']

# Define ARGS with the required parameter
ARGS = {
    'words': words
}
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
euclidean_metric Ran
ran on a synthesized input
asappresearch/dialog-intent-induction/model/utils.py
code served (permissive licence) · get_code("10c7d448434a5383")
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, 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 torch

torch.manual_seed(0)

# Create two small tensors with distinct dimensions
# a: shape (2, 4) - 2 samples, 4-dimensional embeddings
# b: shape (3, 4) - 3 samples, 4-dimensional embeddings
a = torch.randn(2, 4)
b = torch.randn(3, 4)

ARGS = {
    "a": a,
    "b": b,
}
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, 3) — shape (2, 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
get_clusters Ran
ran on a synthesized input
asappresearch/dialog-intent-induction/datasets/askubuntu_preprocess.py
code served (permissive licence) · get_code("68f577055d00b15c")
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 duplicate pairs
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_file.write("615465 8653\n")
temp_file.write("833376 377050\n")
temp_file.write("30585 120621\n")
temp_file.write("178532 152184\n")
temp_file.write("69455 68850\n")
temp_file.write("615465 377050\n")
temp_file.write("833376 120621\n")
temp_file_path = temp_file.name
temp_file.close()

ARGS = {
    'in_dupes_file': temp_file_path,
    'in_max_dupes': None,
    'max_clusters': 3
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
networkx — 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
get_col_by_role Ran
ran on a synthesized input
asappresearch/dialog-intent-induction/labeled_unlabeled_merger.py
code served (permissive licence) · get_code("72d79342f5c121ac")
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

def get_col_by_role(role_string):
    col_by_role = {}
    for s in role_string.split(','):
        split_s = s.split('=')
        col_by_role[split_s[0]] = split_s[1]
    print('col_by_role', col_by_role)
    return col_by_role

ARGS = {
    'role_string': 'feature=col1,target=col2,id=col3'
}
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
get_target Ran
ran on a synthesized input
google/embedding-tests/train_word_embedding_dp.py
code served (permissive licence) · get_code("83f35794b2541212")
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 three parameters:
# - words: a list/sequence that supports indexing and slicing
# - idx: an integer index
# - r: an integer representing the radius/range

# Based on the function body:
# - words needs to support indexing with integers and slicing
# - idx is an integer used as a center point
# - r is an integer used to calculate start and stop bounds

words = ["hello", "world", "foo", "bar", "baz", "qux"]
idx = 3
r = 2

ARGS = {
    "words": words,
    "idx": idx,
    "r": r
}
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
pad_paragraphs Ran
ran on a synthesized input
asappresearch/dialog-intent-induction/model/utils.py
code served (permissive licence) · get_code("bcc6353fd5a3afac")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# Construct a small set of paragraphs for testing
# Each paragraph is a list of sentences, where each sentence is a list of token indices

# Paragraph 1: 2 sentences
para1 = [
    [1, 2, 3],           # sentence 1
    [4, 5]               # sentence 2
]

# Paragraph 2: 3 sentences
para2 = [
    [6, 7, 8, 9],        # sentence 1
    [10, 11],            # sentence 2
    [12]                 # sentence 3
]

# List of paragraphs (batch size 2)
paras = [para1, para2]

ARGS = {
    "paras": paras,
    "pad_idx": 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
pad_sentences Ran
ran on a synthesized input
asappresearch/dialog-intent-induction/model/utils.py
code served (permissive licence) · get_code("c88a716892640c81")
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. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# sents: a list of sentences, where each sentence is a list of token indices
# The function processes lists of integers representing token IDs
sents = [
    [1, 2, 3],
    [4, 5],
]

# lpad: optional left padding token index
lpad = None

# rpad: optional right padding token index
rpad = None

# reverse: whether to reverse each sentence
reverse = False

# pad_idx: the padding token index to use for right-padding to max length
pad_idx = 0

# max_sent_len: maximum length to truncate sentences to
max_sent_len = 100

ARGS = {
    'sents': sents,
    'lpad': lpad,
    'rpad': rpad,
    'reverse': reverse,
    'pad_idx': pad_idx,
    'max_sent_len': max_sent_len,
}
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
sents_to_sparse Ran
ran on a synthesized input
google/embedding-tests/inversion_use.py
code served (permissive licence) · get_code("ab56cab4551b8f35")
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)

# ids is a list of lists of integers (representing token IDs for sentences)
# The function processes this to create a sparse tensor representation
ids = [
    [1, 2, 3],           # sentence 1: 3 tokens
    [4, 5],              # sentence 2: 2 tokens
]

ARGS = {
    "ids": ids
}
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
transform Ran
ran on a synthesized input
asappresearch/dialog-intent-induction/train_qt.py
code served (permissive licence) · get_code("855b1f313f931743")
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 2, dtype family float, shape (70, 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 torch
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

class SimpleEncoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.fc = nn.Linear(4, 3)
    
    def forward(self, x, encoder='v1'):
        return self.fc(x)

model = SimpleEncoder()

data = torch.randn(70, 4)

ARGS = {
    'data': data,
    'model': model
}
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 ndarray, ndim 2, dtype family float, shape (70, 3) — shape (70, 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
filter_labels Not yet run
unverified (harvested, no recorded run)
google/embedding-tests/inversion_bert.py
code served (permissive licence) · get_code("a3f2c032d2c01d6d")
How it ran

no invocation recorded.

filter_sents_indices Not yet run
unverified (harvested, no recorded run)
google/embedding-tests/save_text_embedding.py
code served (permissive licence) · get_code("3012d3468a89592b")
How it ran

no invocation recorded.

get_fetch_by_layer Not yet run
unverified (harvested, no recorded run)
google/embedding-tests/inversion_use.py
code served (permissive licence) · get_code("603d9db381f4b2c1")
How it ran

no invocation recorded.

ids_to_sentence Not yet run
unverified (harvested, no recorded run)
coleridge72/quick_thoughts/analysis.py
pointer only (licence: NONE) · get_code("deda7a319111941b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, coleridge72/quick_thoughts when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

load_posts Not yet run
unverified (harvested, no recorded run)
asappresearch/dialog-intent-induction/datasets/askubuntu_preprocess.py
code served (permissive licence) · get_code("10bf05d8748667ab")
How it ran

no invocation recorded.

mean_pool Not yet run
unverified (harvested, no recorded run)
google/embedding-tests/inversion_use.py
code served (permissive licence) · get_code("69c3a9f159d38fe2")
How it ran

no invocation recorded.

preprocess_raw_data Not yet run
unverified (harvested, no recorded run)
google/embedding-tests/authorship_text.py
code served (permissive licence) · get_code("7585f0e77d8e86db")
How it ran

no invocation recorded.

read_posts_index Not yet run
unverified (harvested, no recorded run)
asappresearch/dialog-intent-induction/datasets/askubuntu_preprocess.py
code served (permissive licence) · get_code("db6c70abdf239846")
How it ran

no invocation recorded.

transform Not yet run
unverified (harvested, no recorded run)
asappresearch/dialog-intent-induction/run_mvsc.py
code served (permissive licence) · get_code("962f3fb3209d777e")
How it ran

no invocation recorded.

Also present in other papers' code

Of 21 samples lifted from this paper's linked repositories, 21 carry a function identity and 0 do not. Of the 21 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 20 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
create_lookup_tables 2 other papers
2 dated earlier
1607.04606 (2016-07)
Enriching Word Vectors with Subword Information
1612.03651 (2016-12)
FastText.zip: Compressing text classification models

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 work we propose a simple and efficient framework for learning sentence representations from unlabelled data. Drawing inspiration from the distributional hypothesis and recent work on learning sentence representations, we reformulate the problem of predicting the context in which a sentence appears as a classification problem. Given a sentence and its context, a classifier distinguishes context sentences from other contrastive sentences based on their vector representations. This allows us to efficiently learn different types of encoding functions, and we show that the model learns high-quality sentence representations. We demonstrate that our sentence representations outperform state-of-the-art unsupervised and supervised representation learning methods on several downstream NLP tasks that involve understanding sentence semantics while achieving an order of magnitude speedup in training time.

For agents

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

get_harvested_code_for_paper("1803.02893")
get_code_for_paper("1803.02893")
have("1803.02893")

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

curl https://syntology.ai/api/ran/1803.02893.json

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

[![Syntology run record](https://syntology.ai/api/ran/1803.02893.svg)](https://syntology.ai/paper/1803.02893)

Connect an agent — have() is free.