SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2402.00838 · ACL · 2024

OLMo : Accelerating the Science of Language Models

Mitchell Wortsman, Oyvind Tafjord, Yizhong Wang, Jack Hessel, Arman Cohan, Kyle Lo, Iz Beltagy, Jesse Dodge, Tushar Khot, Kyle Richardson, William Merrill, Yuling Gu, and 31 more

arXiv · PDF · Open in the Atlas

Code that ran

6 of 13 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 6 have no contract verdict; 0 failures of our instrument; 7 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/2402.00838.json · badge: /api/ran/2402.00838.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
ningyuxu/tip_of_tongue community (classifier label)
classifier basis not recorded on the link
6 ran · 7 no recorded run
FunctionStatusWhere it lives
binarize Ran
ran on a synthesized input
ningyuxu/tip_of_tongue/calf/utils/file.py
code served (permissive licence) · get_code("c445b3401a0c2dc5")
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 os
import mmap
import struct
import pickle
from pathlib import Path
from collections import defaultdict
from typing import Optional, Union, Any, Iterable, Tuple, List, Dict
import torch
import tempfile

torch.manual_seed(0)

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

# Create a simple binary file for testing
test_bin_file = os.path.join(temp_dir, "test.bin")

# First, create a test binary file with some data
test_data = {
    "key1": ["value1", "value2"],
    "key2": ["value3"]
}

# Binarize test data to create a valid binary file
start, meta = (0, defaultdict(list))
with open(test_bin_file, mode='wb') as fw:
    for key, vals in test_data.items():
        for v in vals:
            bv = pickle.dumps(v)
            fw.write(bv)
            meta[key].append((start, len(bv)))
            start = start + len(bv)
    meta_tensor = {key: torch.tensor(val) for key, val in meta.items()}
    pickled = pickle.dumps(meta_tensor)
    fw.write(pickled)
    fw.write(struct.pack('LL', start, len(pickled)))

# Test data for binarize function
data_dict = {
    "samples": ["hello", "world", "test"],
    "labels": [1, 2, 3]
}

output_bin_file = os.path.join(temp_dir, "output.bin")

ARGS = {
    "data": data_dict,
    "bin_file": output_bin_file,
    "merge": False
}
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, 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
extract Ran
ran on a synthesized input
ningyuxu/tip_of_tongue/calf/utils/file.py
code served (permissive licence) · get_code("1b804ad33840c905")
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 tempfile
import zipfile
import gzip
from pathlib import Path

# Create a temporary directory for testing
temp_dir = tempfile.mkdtemp()
temp_path = Path(temp_dir)

# Create a simple zip file for testing
zip_file_path = temp_path / "test.zip"
with zipfile.ZipFile(zip_file_path, 'w') as zf:
    zf.writestr("test_file.txt", "Hello, World!")

# Create a gzip file for testing
gz_file_path = temp_path / "test.txt.gz"
with gzip.open(gz_file_path, 'wb') as f:
    f.write(b"Compressed content")

# Prepare arguments for the extract function
ARGS = {
    "compressed_file": str(zip_file_path),
    "destination_folder": None,
    "clean": False
}
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 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
isfullwidth Ran
ran on a synthesized input
ningyuxu/tip_of_tongue/calf/utils/common.py
code served (permissive licence) · get_code("1b6c470bcbe26474")
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, 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 unicodedata

def isfullwidth(token: str) -> bool:
    return all((unicodedata.east_asian_width(char) in ['W', 'F', 'A'] for char in token))

# Test with various strings
# Full-width characters: Chinese, Japanese, Korean characters are typically 'W' or 'F'
# ASCII characters are typically 'N' (Neutral/Not East Asian)
# Some characters like emoji can be 'A' (Ambiguous)

# Example 1: ASCII string (should be False - all 'N' width)
test_token_1 = "hello"

# Example 2: Full-width string (CJK characters are 'W')
test_token_2 = "你好"

# Example 3: Mixed (should be False)
test_token_3 = "hello世界"

# Example 4: Full-width hiragana (Japanese, typically 'W')
test_token_4 = "ひらがな"

# Example 5: Empty string edge case
test_token_5 = ""

ARGS = {
    "token": test_token_2
}
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, 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
islatin Ran
ran on a synthesized input
ningyuxu/tip_of_tongue/calf/utils/common.py
code served (permissive licence) · get_code("859c276ee0839129")
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, 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 unicodedata

def islatin(token: str) -> bool:
    return all(('LATIN' in unicodedata.name(char) for char in token))

# The function takes a string token and checks if all characters have 'LATIN' in their unicode name
# We need to provide a valid string argument

ARGS = {
    "token": "hello"
}
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, 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
ispunct Ran
ran on a synthesized input
ningyuxu/tip_of_tongue/calf/utils/common.py
code served (permissive licence) · get_code("ef529d3102c72be4")
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, 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 unicodedata

def ispunct(token: str) -> bool:
    return all((unicodedata.category(char).startswith('P') for char in token))

# The function takes a single string parameter and checks if all characters
# are punctuation marks. We need to provide a string value for the token parameter.

ARGS = {
    'token': '!!!'
}
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, 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
wait Ran
ran on a synthesized input
ningyuxu/tip_of_tongue/calf/utils/distributed.py
code served (permissive licence) · get_code("8a3a4a8cc5bb4508")
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 torch
import functools
from typing import Any, List

torch.manual_seed(0)

# The `wait` function takes a single parameter `fn` which is a callable.
# The function wraps `fn` and calls it with *args and **kwargs.
# So we need to provide a simple callable that can be executed.

def dummy_function(x=None, y=None):
    """A simple function to be wrapped by wait decorator"""
    return {"result": 42, "x": x, "y": y}

ARGS = {
    "fn": dummy_function
}
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
format_wordnet_synset Not yet run
unverified (harvested, no recorded run)
ningyuxu/tip_of_tongue/exps/corpus/wordnet.py
code served (permissive licence) · get_code("995b95c68ca8681a")
How it ran

no invocation recorded.

gather Not yet run
unverified (harvested, no recorded run)
ningyuxu/tip_of_tongue/calf/utils/distributed.py
code served (permissive licence) · get_code("f273d6bf63dd6913")
How it ran

no invocation recorded.

get_free_gpus Not yet run
unverified (harvested, no recorded run)
ningyuxu/tip_of_tongue/calf/utils/cuda.py
code served (permissive licence) · get_code("ed5648c3ddfb07b5")
How it ran

no invocation recorded.

get_free_gpus_by_gid Not yet run
unverified (harvested, no recorded run)
ningyuxu/tip_of_tongue/calf/utils/cuda.py
code served (permissive licence) · get_code("b477f3460d3ffc77")
How it ran

no invocation recorded.

reduce Not yet run
unverified (harvested, no recorded run)
ningyuxu/tip_of_tongue/calf/utils/distributed.py
code served (permissive licence) · get_code("a8da323a33fb2f4d")
How it ran

no invocation recorded.

wordnet_synset_data Not yet run
unverified (harvested, no recorded run)
ningyuxu/tip_of_tongue/exps/corpus/wordnet.py
code served (permissive licence) · get_code("1a3fe5eb49125104")
How it ran

no invocation recorded.

wordnet_word_dataset Not yet run
unverified (harvested, no recorded run)
ningyuxu/tip_of_tongue/exps/corpus/wordnet.py
code served (permissive licence) · get_code("904d581818c29bc8")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 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 13 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

Language models (LMs) have become ubiquitous in both NLP research and in commercial product offerings. As their commercial importance has surged, the most powerful models have become closed off, gated behind proprietary interfaces, with important details of their training data, architectures, and development undisclosed. Given the importance of these details in scientifically studying these models, including their biases and potential risks, we believe it is essential for the research community to have access to powerful, truly open LMs. To this end, we have built OLMo, a competitive, truly Open Language Model, to enable the scientific study of language models. Unlike most prior efforts that have only released model weights and inference code, we release OLMo alongside open training data and training and evaluation code. We hope this release will empower the open research community and inspire a new wave of innovation. Yvon, et al. 2022. Bloom: A 176b-parameter openaccess multilingual language model. arXiv preprint arXiv:2211.05100.

For agents

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

get_harvested_code_for_paper("2402.00838")
get_code_for_paper("2402.00838")
have("2402.00838")

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

curl https://syntology.ai/api/ran/2402.00838.json

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

[![Syntology run record](https://syntology.ai/api/ran/2402.00838.svg)](https://syntology.ai/paper/2402.00838)

Connect an agent — have() is free.