SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2211.01786 · 2022

Crosslingual Generalization through Multitask Finetuning

arXiv · PDF · Open in the Atlas

Code that ran

4 of 4 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; 1 failure of our instrument; 0 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/2211.01786.json · badge: /api/ran/2211.01786.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
bigscience-workshop/xmtf canonical (classifier label)
classifier basis not recorded on the link
4 ran · 0 no recorded run
FunctionStatusWhere it lives
apply_template Ran
ran on a synthesized input
bigscience-workshop/xmtf/create_xp3x.py
code served (permissive licence) · get_code("3cea2e48e72663a4")
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 MockDataset. 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
from unittest.mock import Mock

torch.manual_seed(0)

# Create a minimal template object that has an apply method
class MockTemplate:
    def apply(self, ex, strip_connection=True, truncate=True):
        # Returns a tuple of (inputs_str, targets_list)
        inputs = f"Input based on {ex.get('text', 'default')}"
        targets = ["target_output"]
        return inputs, targets

# Create a minimal dataset-like object
class MockDataset:
    def __init__(self):
        self.column_names = ['text', 'label', 'extra-field']
        self.data = [
            {'text': 'sample one', 'label': 'A', 'extra-field': 'value1'},
            {'text': 'sample two', 'label': 'B', 'extra-field': 'value2'},
        ]
        self.current_idx = 0
    
    def map(self, fn):
        new_data = [fn(ex) for ex in self.data]
        dataset = MockDataset()
        dataset.data = new_data
        dataset.column_names = list(new_data[0].keys()) if new_data else []
        return dataset
    
    def filter(self, fn):
        new_data = [ex for ex in self.data if fn(ex)]
        dataset = MockDataset()
        dataset.data = new_data
        dataset.column_names = self.column_names
        return dataset
    
    def remove_columns(self, cols_to_remove):
        if not cols_to_remove:
            return self
        new_data = []
        for ex in self.data:
            new_ex = {k: v for k, v in ex.items() if k not in cols_to_remove}
            new_data.append(new_ex)
        dataset = MockDataset()
        dataset.data = new_data
        dataset.column_names = [c for c in self.column_names if c not in cols_to_remove]
        return dataset

# Create test data
template = MockTemplate()
dataset = MockDataset()

ARGS = {
    'dataset': dataset,
    'template': template,
    'strip_connection': True,
}
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 MockDataset — a MockDataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MockDataset 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_lev Ran
ran on a synthesized input
bigscience-workshop/xmtf/plotstables/levenshtein.py
code served (permissive licence) · get_code("b548f6ca8e44e7dc")
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
import numpy as np
from Levenshtein import distance as levenshtein_distance

torch.manual_seed(0)

# Create a sample dict with 'premise' and 'hypothesis' string keys
sample = {
    'premise': 'The quick brown fox',
    'hypothesis': 'The quick brown dog'
}

ARGS = {
    'sample': sample
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
Levenshtein, numpy, 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
count_comments Ran
ran on a synthesized input
bigscience-workshop/xmtf/plotstables/countcode.py
code served (permissive licence) · get_code("e3b3dd5f405fc865")
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

torch.manual_seed(0)

ARGS = {
    'x': {
        'targets': '#hello#world#test#',
        'counts': 0
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
removeHyphen Ran
ran; our contract draft was wrong, not the code
bigscience-workshop/xmtf/create_xp3x.py
code served (permissive licence) · get_code("0b1cb631e6e315b0")
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

torch.manual_seed(0)

ARGS = {
    "example": {
        "key-with-hyphen": "value1",
        "normal_key": "value2",
        "another-hyphenated-key": 42,
        "no_hyphen": [1, 2, 3]
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 1 later) and 3 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
removeHyphen 3 other papers
1 dated earlier, 2 later
2110.08207 (2021-10)
Multitask Prompted Training Enables Zero-Shot Task Generalization
2310.20707 (2023-10)
What's In My Big Data?
2408.03617 (2024-08)
Is Child-Directed Speech Effective Training Data for Language Models?

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

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

Abstract

Multitask prompted finetuning (MTF) has been shown to help large language models generalize to new tasks in a zero-shot setting, but so far explorations of MTF have focused on English data and models. We apply MTF to the pretrained multilingual BLOOM and mT5 model families to produce finetuned variants called BLOOMZ and mT0. We find finetuning large multilingual language models on English tasks with English prompts allows for task generalization to non-English languages that appear only in the pretraining corpus. Finetuning on multilingual tasks with English prompts further improves performance on English and non-English tasks leading to various state-of-the-art zero-shot results. We also investigate finetuning on multilingual tasks with prompts that have been machine-translated from English to match the language of each dataset. We find training on these machine-translated prompts leads to better performance on human-written prompts in the respective languages. Surprisingly, we find models are capable of zero-shot generalization to tasks in languages they have never intentionally seen. We conjecture that the models are learning higher-level capabilities that are both task- and language-agnostic. In addition, we introduce xP3, a composite of supervised datasets in 46 languages with English and machine-translated prompts. Our code, datasets and models are freely available at https://github.com/bigscience-workshop/xmtf.

For agents

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

get_harvested_code_for_paper("2211.01786")
get_code_for_paper("2211.01786")
have("2211.01786")

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

curl https://syntology.ai/api/ran/2211.01786.json

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

[![Syntology run record](https://syntology.ai/api/ran/2211.01786.svg)](https://syntology.ai/paper/2211.01786)

Connect an agent — have() is free.