SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2204.08582 · 2022

MASSIVE: A 1M-Example Multilingual Natural Language Understanding Dataset with 51 Typologically-Diverse Languages

arXiv · PDF · Open in the Atlas

Code that ran

8 of 10 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, 1 violated it and 7 have no contract verdict; 0 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/2204.08582.json · badge: /api/ran/2204.08582.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
alexa/massive canonical (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
rita-nlp/italic community (classifier label)
classifier basis not recorded on the link
7 ran · 2 no recorded run
FunctionStatusWhere it lives
compute_metrics Ran
ran on a synthesized input
rita-nlp/italic/text_finetuning.py
code served (permissive licence) · get_code("652b541e7a5feac9")
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 numpy as np
from sklearn.metrics import accuracy_score

# The function expects eval_pred to be a tuple of (predictions, labels)
# predictions should be either:
#   - a 2D array of shape (batch_size, num_classes) with logits/probabilities
#   - OR a tuple/list where predictions[0] is such an array
# labels should be a 1D array of true class indices

# Create predictions: logits for 2 samples, 4 classes
predictions = np.random.randn(2, 4).astype(np.float32)

# Create labels: true class indices for 2 samples
labels = np.array([0, 2], dtype=np.int64)

# Construct the eval_pred tuple
eval_pred = (predictions, labels)

ARGS = {
    'eval_pred': eval_pred
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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
compute_metrics Ran
ran on a synthesized input
rita-nlp/italic/text_inference.py
code served (permissive licence) · get_code("0a2872c4352696b2")
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 numpy as np

# The function expects eval_pred which is a tuple of (predictions, labels)
# predictions can be either:
#   1. A 2D array where argmax is taken along axis=1
#   2. A tuple/list where predictions[0] is a 2D array
# labels should be a 1D array of true labels

# Create sample predictions (logits from a classifier)
# Shape: (batch_size=2, num_classes=4)
predictions = np.random.randn(2, 4).astype(np.float32)

# Create sample labels (true class indices)
# Shape: (batch_size=2,)
labels = np.array([0, 2], dtype=np.int64)

# eval_pred is a tuple of (predictions, labels)
eval_pred = (predictions, labels)

ARGS = {
    'eval_pred': eval_pred
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — 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
compute_metrics Ran
ran on a synthesized input
rita-nlp/italic/utils.py
code served (permissive licence) · get_code("f3b2d10b53645633")
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 numpy as np
import torch

# Create a simple object to hold predictions and labels
class PredictionOutput:
    def __init__(self, predictions, label_ids):
        self.predictions = predictions
        self.label_ids = label_ids

# Set seed for reproducibility
torch.manual_seed(0)
np.random.seed(0)

# Create small prediction outputs
# 4 samples, 3 classes
batch_size = 4
num_classes = 3

predictions = np.random.randn(batch_size, num_classes).astype(np.float32)
label_ids = np.array([0, 1, 2, 1], dtype=np.int64)

pred = PredictionOutput(predictions=predictions, label_ids=label_ids)

ARGS = {
    'pred': pred
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn, 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_text Ran
ran on a synthesized input
rita-nlp/italic/asr_inference.py
code served (permissive licence) · get_code("fef1c66dfeeb6664")
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

torch.manual_seed(0)

# The function get_text accesses sample['utt']
# So we need a dict with a 'utt' key
sample = {
    'utt': 'hello world'
}

ARGS = {
    'sample': sample
}
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
is_audio_in_length_range Ran
ran on a synthesized input
rita-nlp/italic/asr_finetuning.py
code served (permissive licence) · get_code("0a219dbdc5c3d3aa")
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 torch
import argparse

# Set up the args namespace that parse_args() would return
class Args:
    def __init__(self):
        self.model_name_or_path = 'EdoAbati/whisper-medium-it-2'
        self.dataset_name_or_path = 'RiTA-nlp/italic-easy'
        self.batch_size = 8
        self.learning_rate = 1e-05
        self.num_train_epochs = 5
        self.max_input_length_in_seconds = 15.0
        self.gradient_accumulation_steps = 1

args = Args()

def is_audio_in_length_range(length):
    return length < args.max_input_length_in_seconds

# Define ARGS dict with test values for the function parameter
ARGS = {
    'length': 10.5
}
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 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
is_target_text_in_range Ran
ran on a synthesized input
rita-nlp/italic/asr_inference.py
code served (permissive licence) · get_code("b1dc84afdc06af86")
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 torch

torch.manual_seed(0)

ARGS = {
    'ref': 'ignore time segment in scoring'
}
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 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
isascii Ran
ran, violated the contract we drafted
alexa/massive/scripts/create_hf_dataset.py
pointer only (licence: NOASSERTION) · get_code("00de6205350da403")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, alexa/massive (NOASSERTION). 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. A run is recorded.

normalise Ran
ran on a synthesized input
rita-nlp/italic/asr_inference.py
code served (permissive licence) · get_code("48422a4b6a62ca2f")
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
from transformers.models.whisper.english_normalizer import BasicTextNormalizer

torch.manual_seed(0)

# The function 'normalise' takes a single parameter 'batch'
# Looking at the function body:
# - batch is accessed like a dict with keys 'utt' and 'norm_text'
# - get_text(batch) calls batch['utt']
# - whisper_norm() is called on the result of get_text(batch)
# - The result is stored back in batch['norm_text']

# So batch needs to be a dict-like object with at least an 'utt' key
# containing text that can be normalized

batch = {
    'utt': 'hello world THIS IS A TEST',
}

ARGS = {
    'batch': batch
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch, transformers — 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
define_model Not yet run
unverified (harvested, no recorded run)
rita-nlp/italic/ic_finetuning.py
code served (permissive licence) · get_code("53d7e4178b94a91c")
How it ran

no invocation recorded.

define_model Not yet run
unverified (harvested, no recorded run)
rita-nlp/italic/ic_inference.py
code served (permissive licence) · get_code("d253d61bd1e78474")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 9 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 9 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

We present the MASSIVE dataset--Multilingual Amazon Slu resource package (SLURP) for Slot-filling, Intent classification, and Virtual assistant Evaluation. MASSIVE contains 1M realistic, parallel, labeled virtual assistant utterances spanning 51 languages, 18 domains, 60 intents, and 55 slots. MASSIVE was created by tasking professional translators to localize the English-only SLURP dataset into 50 typologically diverse languages from 29 genera. We also present modeling results on XLM-R and mT5, including exact match accuracy, intent classification accuracy, and slot-filling F1 score. We have released our dataset, modeling code, and models publicly.

For agents

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

get_harvested_code_for_paper("2204.08582")
get_code_for_paper("2204.08582")
have("2204.08582")

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

curl https://syntology.ai/api/ran/2204.08582.json

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

[![Syntology run record](https://syntology.ai/api/ran/2204.08582.svg)](https://syntology.ai/paper/2204.08582)

Connect an agent — have() is free.