SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2109.12072 · 2021

SD-QA: Spoken Dialectal Question Answering for the Real World

arXiv · PDF · Open in the Atlas

Code that ran

20 of 22 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 17 have no contract verdict; 3 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/2109.12072.json · badge: /api/ran/2109.12072.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
ffaisal93/sd-qa canonical (classifier label)
classifier basis: census
20 ran · 2 no recorded run
FunctionStatusWhere it lives
byte_len Ran
ran on a synthesized input
ffaisal93/sd-qa/baselines/tydiqa/baseline/data.py
code served (permissive licence) · get_code("7671754d0cee7d11")
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 int, ndim 0, dtype family int, 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 torch
import numpy as np

torch.manual_seed(0)

def byte_str(text):
    return text.encode('utf-8')

def byte_len(text):
    return len(byte_str(text))

# The function byte_len takes a single parameter 'text' which should be a string
# It encodes the string to UTF-8 bytes and returns the length

ARGS = {
    'text': 'Hello, world!'
}
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 int, ndim 0, dtype family int, 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
byte_slice Ran
ran on a synthesized input
ffaisal93/sd-qa/baselines/tydiqa/tydi_eval.py
code served (permissive licence) · get_code("787fbf63d8c1ab07")
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
import numpy as np

# Analyze the function:
# - text: a string that will be encoded to UTF-8 bytes
# - start: integer index for byte slicing
# - end: integer index for byte slicing
# The function encodes text to UTF-8 bytes, slices from start to end,
# and attempts to decode back to string (or returns the bytes repr if decode fails)

ARGS = {
    "text": "Hello, 世界!",
    "start": 0,
    "end": 5
}
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 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
byte_slice Ran
ran on a synthesized input
ffaisal93/sd-qa/baselines/tydiqa/baseline/data.py
code served (permissive licence) · get_code("5e27cb5b15ab895f")
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
import numpy as np

torch.manual_seed(0)

# Analyzing the function byte_slice:
# - text: a string that will be encoded to utf-8 bytes, sliced, then decoded
# - start: integer index for slice start
# - end: integer index for slice end
# - errors: string parameter for decode error handling (default 'replace')

ARGS = {
    'text': 'Hello, World! 🌍',
    'start': 0,
    'end': 13,
    'errors': 'replace'
}
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 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
byte_str Ran
ran on a synthesized input
ffaisal93/sd-qa/baselines/tydiqa/baseline/data.py
code served (permissive licence) · get_code("a5845e18acb8844b")
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 bytes. 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)

ARGS = {
    "text": "Hello, world!"
}
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 bytes — a bytes output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a bytes 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
dir_check Ran
ran on a synthesized input
ffaisal93/sd-qa/experiments/utils.py
code served (permissive licence) · get_code("ea2aa48fb12d0a30")
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 os
import tempfile

# Create a temporary directory for testing
with tempfile.TemporaryDirectory() as tmpdir:
    # Create a test directory
    test_dir = os.path.join(tmpdir, "test_directory")
    os.makedirs(test_dir, exist_ok=True)
    
    # Create a test file
    test_file = os.path.join(tmpdir, "test_file.txt")
    with open(test_file, 'w') as f:
        f.write("test content")
    
    # Define arguments for both test cases
    ARGS = {
        'path': test_dir,
        'types': 'dir'
    }
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
extract_answers Ran
ran on a synthesized input
ffaisal93/sd-qa/extract_qa_from_jsonl.py
code served (permissive licence) · get_code("fd24ba8d7375d425")
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
import numpy as np

ARGS = {
    'example': {
        'document_plaintext': 'The quick brown fox jumps over the lazy dog. Python is great.',
        'annotations': [
            {
                'minimal_answer': {
                    'plaintext_start_byte': 4,
                    'plaintext_end_byte': 9
                },
                'passage_answer': {
                    'candidate_index': -1
                }
            },
            {
                'minimal_answer': {
                    'plaintext_start_byte': -1,
                    'plaintext_end_byte': -1
                },
                'passage_answer': {
                    'candidate_index': 0
                }
            },
            {
                'minimal_answer': None,
                'passage_answer': {
                    'candidate_index': 1
                }
            }
        ],
        'passage_answer_candidates': [
            {
                'plaintext_start_byte': 16,
                'plaintext_end_byte': 19
            },
            {
                'plaintext_start_byte': 48,
                'plaintext_end_byte': 54
            }
        ]
    }
}
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 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
gelu Ran
ran; our contract draft was wrong, not the code
ffaisal93/sd-qa/baselines/tydiqa/baseline/bert/modeling.py
code served (permissive licence) · get_code("51d3bc0cfa6415a3")
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 SymbolicTensor. 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 numpy as np
import tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

# Create a small float32 tensor for x
x = tf.constant([[0.5, -0.3, 1.2, -0.8],
                  [0.1, 0.2, -0.5, 0.9]], dtype=tf.float32)

ARGS = {
    "x": x
}
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 SymbolicTensor — a SymbolicTensor 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_activation Ran
ran; our contract draft was wrong, not the code
ffaisal93/sd-qa/baselines/tydiqa/baseline/bert/modeling.py
code served (permissive licence) · get_code("217fe96a5945a42d")
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 numpy as np
import six
import tensorflow.compat.v1 as tf

tf.disable_v2_behavior()

ARGS = {
    'activation_string': 'relu'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, six, 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
get_assignment_map_from_checkpoint Ran
ran; our contract draft was wrong, not the code
ffaisal93/sd-qa/baselines/tydiqa/baseline/bert/modeling.py
code served (permissive licence) · get_code("7bc46f4950599a87")
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 collections
import re
import tensorflow.compat.v1 as tf
import tempfile
import os

# Disable eager execution for TensorFlow v1 compatibility
tf.disable_eager_execution()

# Create a temporary checkpoint file for testing
temp_dir = tempfile.mkdtemp()
checkpoint_path = os.path.join(temp_dir, "test_checkpoint")

# Create a simple TensorFlow checkpoint with some variables
with tf.Graph().as_default():
    with tf.Session() as sess:
        # Create some variables to save
        var1 = tf.Variable(tf.ones([2, 3]), name="layer1/weights")
        var2 = tf.Variable(tf.zeros([4]), name="layer1/bias")
        var3 = tf.Variable(tf.ones([3, 4]), name="layer2/weights")
        
        sess.run(tf.global_variables_initializer())
        
        # Save the checkpoint
        saver = tf.train.Saver()
        saver.save(sess, checkpoint_path)

# Create variables for the current graph
with tf.Graph().as_default():
    with tf.Session() as sess:
        # Create variables with matching names
        tvars = [
            tf.Variable(tf.ones([2, 3]), name="layer1/weights"),
            tf.Variable(tf.zeros([4]), name="layer1/bias"),
            tf.Variable(tf.ones([3, 4]), name="layer2/weights"),
            tf.Variable(tf.ones([5]), name="layer3/bias"),  # This one won't be in checkpoint
        ]
        
        sess.run(tf.global_variables_initializer())
        
        ARGS = {
            "tvars": tvars,
            "init_checkpoint": checkpoint_path,
        }
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 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
get_metadata Ran
ran on a synthesized input
ffaisal93/sd-qa/experiments/utils.py
code served (permissive licence) · get_code("d60c3869c082ce69")
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 pandas as pd
import tempfile
import shutil

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

# Create the directory structure: root_path/f1/language/subdirs
f1_name = "test_split"
f1_path = os.path.join(temp_dir, f1_name)
os.makedirs(f1_path)

# Create language directories and CSV files
languages = ['ben', 'eng', 'ara', 'kor', 'swa']
lang_full_names = {'ben': 'bengali', 'eng': 'english', 'ara': 'arabic', 'kor': 'korean', 'swa': 'swahili'}

for lang in languages:
    lang_path = os.path.join(f1_path, lang)
    os.makedirs(lang_path)
    
    # Create the language CSV file
    csv_filename = os.path.join(lang_path, f"{lang_full_names[lang]}.{f1_name}.csv")
    
    # Create sample data
    data = {
        'Unnamed: 0': [0, 1],
        f"{lang_full_names[lang]}-gold": ['word1 ', 'word2'],
    }
    
    # Add kenya-en-KE column for eng
    if lang == 'eng':
        data['kenya-en-KE'] = ['value1', 'value2']
    else:
        data['kenya-en-KE'] = ['#NOTFOUND#', '#NOTFOUND#']
    
    df = pd.DataFrame(data)
    df.to_csv(csv_filename)
    
    # Create subdirectories with metadata files
    for subdir_idx in range(2):
        subdir_name = f"subset_{subdir_idx}"
        subdir_path = os.path.join(lang_path, subdir_name)
        os.makedirs(subdir_path)
        
        # Create metadata CSV file
        meta_filename = os.path.join(subdir_path, "metadata.csv")
        meta_data = {
            'Unnamed: 0': [0, 1],
            'language': ['swh' if lang == 'swa' else lang, 'swh' if lang == 'swa' else lang],
            'other_field': ['val1', 'val2']
        }
        meta_df = pd.DataFrame(meta_data)
        meta_df.to_csv(meta_filename)

# Define discard set
discard = set()

ARGS = {
    'root_path': temp_dir,
    'f1': f1_name,
    'discard': discard
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — 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
load_jsonl_example_ids Ran
ran on a synthesized input
ffaisal93/sd-qa/analyze_data_links.py
code served (permissive licence) · get_code("da7b79cb32612610")
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
from pathlib import Path

# Create a temporary directory for test files
temp_dir = Path(tempfile.gettempdir())

# Create a test JSONL file with example_ids
test_jsonl_file = temp_dir / "test_examples.jsonl"
with open(test_jsonl_file, 'w', encoding='utf-8') as f:
    f.write(json.dumps({'example_id': 'ex_001', 'data': 'sample'}) + '\n')
    f.write(json.dumps({'example_id': 'ex_002', 'data': 'sample2'}) + '\n')
    f.write(json.dumps({'example_id': 123, 'data': 'sample3'}) + '\n')
    f.write('\n')  # Empty line to test filtering
    f.write(json.dumps({'data': 'no_id'}) + '\n')

ARGS = {
    'jsonl_file': test_jsonl_file,
}
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
load_metadata_ids Ran
ran on a synthesized input
ffaisal93/sd-qa/analyze_data_links.py
code served (permissive licence) · get_code("50ed7d3089187d7d")
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 set. 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 csv
from pathlib import Path
import tempfile

# Create a temporary directory and metadata file for testing
temp_dir = tempfile.TemporaryDirectory()
temp_path = Path(temp_dir.name)
metadata_file = temp_path / "metadata.csv"

# Write a sample metadata.csv file with some example_ids
with open(metadata_file, 'w', newline='', encoding='utf-8') as f:
    writer = csv.DictWriter(f, fieldnames=['example_id', 'other_field'])
    writer.writeheader()
    writer.writerow({'example_id': 'id_001', 'other_field': 'value1'})
    writer.writerow({'example_id': 'id_002', 'other_field': 'value2'})
    writer.writerow({'example_id': '  id_003  ', 'other_field': 'value3'})
    writer.writerow({'example_id': '', 'other_field': 'value4'})

ARGS = {
    'metadata_file': metadata_file
}
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 set — a set output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a set 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
load_txt_ids Ran
ran on a synthesized input
ffaisal93/sd-qa/analyze_data_links.py
code served (permissive licence) · get_code("5d1dc7f1ffae5ff9")
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 set. 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.

from pathlib import Path
import tempfile
import os

# Create a temporary txt file with tab-separated data
temp_dir = tempfile.mkdtemp()
txt_file_path = Path(temp_dir) / "test_ids.txt"

# Write test data to the file
with open(txt_file_path, 'w', encoding='utf-8') as f:
    f.write("id1\tvalue1\tvalue2\n")
    f.write("id2\tvalue3\tvalue4\n")
    f.write("id3\tvalue5\tvalue6\n")

ARGS = {
    "txt_file": txt_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 set — a set output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a set 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
read_reference Ran
ran on a synthesized input
ffaisal93/sd-qa/experiments/transcript_replace.py
code served (permissive licence) · get_code("91c7ebc9939cc186")
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 tempfile
import csv
import os

# Create a temporary CSV file with test data
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.csv', newline='')
temp_path = temp_file.name

try:
    writer = csv.writer(temp_file)
    writer.writerow(['ID', 'question', 'other_col'])
    writer.writerow(['1', 'What is this', 'data'])
    writer.writerow(['2', 'How are you', 'more'])
    writer.writerow(['3', 'Is this working', 'stuff'])
    temp_file.close()
    
    ARGS = {
        'f': temp_path,
        'col': 'question',
        'qmark': None
    }
finally:
    pass
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
safe_average Ran
ran on a synthesized input
ffaisal93/sd-qa/baselines/tydiqa/eval_utils.py
code served (permissive licence) · get_code("e5609b0f6acf5d82")
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 float, 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 torch
import numpy as np

torch.manual_seed(0)

# safe_average takes a single parameter: elements
# It calls sum(elements) and len(elements), so elements must be an iterable
# Let's create a list of numbers

elements = [1.0, 2.0, 3.0, 4.0, 5.0]

ARGS = {
    "elements": elements
}
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 float, 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
safe_divide Ran
ran on a synthesized input
ffaisal93/sd-qa/baselines/tydiqa/eval_utils.py
code served (permissive licence) · get_code("999c2750713a6692")
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 float, 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 torch

torch.manual_seed(0)

ARGS = {
    "x": 10,
    "y": 2,
}
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 float, 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
sanitize_field Ran
ran on a synthesized input
ffaisal93/sd-qa/extract_qa_from_jsonl.py
code served (permissive licence) · get_code("606f24554fa9223b")
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
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "s": "hello\tworld\nfoo\r\nbar  baz"
}
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
utf8_slice Ran
ran on a synthesized input
ffaisal93/sd-qa/extract_qa_from_jsonl.py
code served (permissive licence) · get_code("812e406d9764d77d")
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
import numpy as np

# Analyze the function utf8_slice:
# - text: str - a string to slice
# - start: int - byte offset for start (inclusive)
# - end: int - byte offset for end (exclusive)
# The function slices UTF-8 encoded bytes and decodes back to string

# Create test arguments
text_value = "Hello, 世界! This is UTF-8 text."
start_value = 0
end_value = 13

ARGS = {
    "text": text_value,
    "start": start_value,
    "end": end_value,
}
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 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
utf8_slice Ran
ran on a synthesized input
ffaisal93/sd-qa/sdqa_triples_from_jsonl.py
code served (permissive licence) · get_code("f4a930be2113e081")
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
import numpy as np

# Construct arguments for utf8_slice function
# The function takes: text (str), start (int), end (int)

text = "Hello, 世界! 🌍"  # Mixed ASCII and Unicode characters
start = 0
end = 20

ARGS = {
    "text": text,
    "start": start,
    "end": end
}
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 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
write_out Ran
ran on a synthesized input
ffaisal93/sd-qa/experiments/transcript_replace.py
code served (permissive licence) · get_code("dad8041c5ab85e66")
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 json
import torch
import numpy as np

torch.manual_seed(0)

# The function write_out takes a single parameter 'line' and converts it to JSON
# It uses json.dumps with ensure_ascii=False
# So 'line' should be a JSON-serializable object

# Test with various types that json.dumps can handle
test_line = {
    "text": "Hello, world!",
    "number": 42,
    "float": 3.14,
    "list": [1, 2, 3],
    "nested": {"key": "value"},
    "unicode": "こんにちは"
}

ARGS = {
    "line": test_line
}
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 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
compute_partial_match_scores Not yet run
unverified (harvested, no recorded run)
ffaisal93/sd-qa/baselines/tydiqa/eval_utils.py
code served (permissive licence) · get_code("7ef49e1793449f8b")
How it ran

no invocation recorded.

find_audio_dir Not yet run
unverified (harvested, no recorded run)
ffaisal93/sd-qa/sdqa_triples_from_jsonl.py
code served (permissive licence) · get_code("d19afd0251c48ecd")
How it ran

no invocation recorded.

Also present in other papers' code

Of 22 samples lifted from this paper's linked repositories, 22 carry a function identity and 0 do not. Of the 22 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 3 later) and 19 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
get_activation 32 other papers
24 dated earlier, 8 later · 1 only as an assembly brick
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.04085 (2019-01)
Passage Re-ranking with BERT
and 29 more not named here
gelu 26 other papers
19 dated earlier, 7 later · 4 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
and 23 more not named here
get_assignment_map_from_checkpoint 26 other papers
20 dated earlier, 6 later
1804.05392 (2018-04)
Higher-order Coreference Resolution with Coarse-to-fine Inference
1808.09781 (2018-08)
Self-Attentive Sequential Recommendation
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
and 23 more not named here

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

Question answering (QA) systems are now available through numerous commercial applications for a wide variety of domains, serving millions of users that interact with them via speech interfaces. However, current benchmarks in QA research do not account for the errors that speech recognition models might introduce, nor do they consider the language variations (dialects) of the users. To address this gap, we augment an existing QA dataset to construct a multi-dialect, spoken QA benchmark on five languages (Arabic, Bengali, English, Kiswahili, Korean) with more than 68k audio prompts in 24 dialects from 255 speakers. We provide baseline results showcasing the real-world performance of QA systems and analyze the effect of language variety and other sensitive speaker attributes on downstream performance. Last, we study the fairness of the ASR and QA models with respect to the underlying user populations. The dataset, model outputs, and code for reproducing all our experiments are available: https://github.com/ffaisal93/SD-QA.

For agents

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

get_harvested_code_for_paper("2109.12072")
get_code_for_paper("2109.12072")
have("2109.12072")

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

curl https://syntology.ai/api/ran/2109.12072.json

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

[![Syntology run record](https://syntology.ai/api/ran/2109.12072.svg)](https://syntology.ai/paper/2109.12072)

Connect an agent — have() is free.