SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.07540 · EMNLP · 2021

Generating Datasets with Pretrained Language Models

Hinrich Schütze, Timo Schick

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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 3 have no contract verdict; 0 failures of our instrument; 0 no recorded run), measured 2026-11-21; 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/2104.07540.json · badge: /api/ran/2104.07540.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
yipingnus/scratchplot-story-generation community (classifier label)
classifier basis not recorded on the link
3 ran · 0 no recorded run
FunctionStatusWhere it lives
read_inputs Ran
ran on a synthesized input
yipingnus/scratchplot-story-generation/utils.py
code served (permissive licence) · get_code("003c442d9c01a497")
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 tempfile
import os
import json
import csv

# Create temporary files for testing
temp_dir = tempfile.mkdtemp()

# Create a plain text file
plain_text_path = os.path.join(temp_dir, "plain.txt")
with open(plain_text_path, 'w', encoding='utf8') as f:
    f.write("This is the first input\n")
    f.write("This is the second input\n")

# Create a JSONL file
jsonl_path = os.path.join(temp_dir, "data.jsonl")
with open(jsonl_path, 'w', encoding='utf8') as f:
    f.write(json.dumps({"text_a": "First text", "text_b": None, "label": 0}) + "\n")
    f.write(json.dumps({"text_a": "Second text", "text_b": None, "label": 1}) + "\n")

# Create an STS benchmark TSV file
stsb_path = os.path.join(temp_dir, "sts.tsv")
with open(stsb_path, 'w', encoding='utf8', newline='') as f:
    writer = csv.writer(f, delimiter='\t', quoting=csv.QUOTE_NONE)
    # STS format: genre year split score sent_a sent_b
    writer.writerow(["news", "2012", "test", "5.00", "dummy1", "Sentence A", "Sentence B"])
    writer.writerow(["news", "2012", "test", "4.50", "dummy2", "Another A", "Another B"])

# Test with plain text file
ARGS = {
    "input_file": plain_text_path,
    "input_file_type": "plain"
}
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
read_jsonl_inputs Ran
ran on a synthesized input
yipingnus/scratchplot-story-generation/utils.py
code served (permissive licence) · get_code("ba94c4cb5a503946")
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 json
import tempfile
import os
from typing import List, Optional, Any

class DatasetEntry:
    """This class represents a dataset entry for text (pair) classification"""

    def __init__(self, text_a: str, text_b: Optional[str] = None, label: Any = None):
        self.text_a = text_a
        self.text_b = text_b
        self.label = label

    def __repr__(self):
        if self.text_b is not None:
            return f'DatasetEntry(text_a="{self.text_a}", text_b="{self.text_b}", label={self.label})'
        else:
            return f'DatasetEntry(text_a="{self.text_a}", label={self.label})'

    def __key(self):
        return (self.text_a, self.text_b, self.label)

    def __hash__(self):
        return hash(self.__key())

    def __eq__(self, other):
        if isinstance(other, DatasetEntry):
            return self.__key() == other.__key()
        return False

    @staticmethod
    def save_list(entries: List['DatasetEntry'], path: str):
        with open(path, 'w', encoding='utf8') as fh:
            for entry in entries:
                fh.write(f'{json.dumps(entry.__dict__)}\n')

    @staticmethod
    def read_list(path: str) -> List['DatasetEntry']:
        pairs = []
        with open(path, 'r', encoding='utf8') as fh:
            for line in fh:
                pairs.append(DatasetEntry(**json.loads(line)))
        return pairs


# Create a temporary JSONL file with test data
temp_dir = tempfile.mkdtemp()
test_file_path = os.path.join(temp_dir, "test_inputs.jsonl")

# Create sample dataset entries
entries = [
    DatasetEntry(text_a="Hello world", text_b=None, label=0),
    DatasetEntry(text_a="This is a test", text_b="Another text", label=1),
    DatasetEntry(text_a="Sample input", text_b=None, label=2),
]

# Save to JSONL file
DatasetEntry.save_list(entries, test_file_path)

ARGS = {
    "path": test_file_path
}
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
read_plaintext_inputs Ran
ran on a synthesized input
yipingnus/scratchplot-story-generation/utils.py
code served (permissive licence) · get_code("a8784da93958c3c8")
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 tempfile
import os

# Create a temporary file with sample plaintext inputs
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt', encoding='utf8')
temp_file.write("This is the first input line\n")
temp_file.write("This is the second input line\n")
temp_file.write("This is the third input line\n")
temp_file.close()

ARGS = {
    'path': 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 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 0 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
read_inputs 1 other paper
1 later
2210.16798 (2022-10)
Generate, Discriminate and Contrast: A Semi-Supervised Sentence Representation Learning Framework
read_jsonl_inputs 1 other paper
1 later
2210.16798 (2022-10)
Generate, Discriminate and Contrast: A Semi-Supervised Sentence Representation Learning Framework
read_plaintext_inputs 1 other paper
1 later
2210.16798 (2022-10)
Generate, Discriminate and Contrast: A Semi-Supervised Sentence Representation Learning Framework

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 2026-11-21.

Repositories linked to this paper

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

Abstract

To obtain high-quality sentence embeddings from pretrained language models (PLMs), they must either be augmented with additional pretraining objectives or finetuned on a large set of labeled text pairs. While the latter approach typically outperforms the former, it requires great human effort to generate suitable datasets of sufficient size. In this paper, we show how PLMs can be leveraged to obtain high-quality sentence embeddings without the need for labeled data, finetuning or modifications to the pretraining objective: We utilize the generative abilities of large and high-performing PLMs to generate entire datasets of labeled text pairs from scratch, which we then use for finetuning much smaller and more efficient models. Our fully unsupervised approach outperforms strong baselines on several semantic textual similarity datasets. 1

For agents

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

get_harvested_code_for_paper("2104.07540")
get_code_for_paper("2104.07540")
have("2104.07540")

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

curl https://syntology.ai/api/ran/2104.07540.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.07540.svg)](https://syntology.ai/paper/2104.07540)

Connect an agent — have() is free.