SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1806.08730 · 2018

The Natural Language Decathlon: Multitask Learning as Question Answering

arXiv · PDF · Open in the Atlas

Code that ran

13 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 12 have no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-21; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/1806.08730.json · badge: /api/ran/1806.08730.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
salesforce/decaNLP canonical (classifier label)
classifier basis: census
9 ran · 0 no recorded run
eric-erki/decaNLP community (classifier label)
classifier basis: plan_row
4 ran · 0 no recorded run
FunctionStatusWhere it lives
all_reverse Ran
ran on a synthesized input
salesforce/decaNLP/validate.py
code served (permissive licence) · get_code("1d88554176b7eac1")
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 Tensor, ndim 2, dtype family int, shape (2, 4) 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 torch.nn as nn

torch.manual_seed(0)

# Create a tiny tensor (batch=2, dim=4)
tensor = torch.randint(0, 10, (2, 4))

# world_size parameter - use 1 to avoid distributed operations
world_size = 1

# Create a field object with a reverse method
class Field:
    def reverse(self, tensor):
        # reverse() should return a tensor that can be sliced with [:clip]
        return torch.flip(tensor, dims=[0])

field = Field()

# clip parameter - clips the result to first N elements
clip = 3

ARGS = {
    "tensor": tensor,
    "world_size": world_size,
    "field": field,
    "clip": clip,
    "dim": 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 Tensor, ndim 2, dtype family int, shape (2, 4) — shape (2, 4) 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
correct_format Ran
ran on a synthesized input
salesforce/decaNLP/convert_to_logical_forms.py
code served (permissive licence) · get_code("b35cd8b933ae1b45")
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)

# The function expects a dict x with:
# - keys() method (dict has this)
# - 'conds' key that contains an iterable of items
# - each item in conds should have len() >= 3 to be "proper"

# Valid case: dict with 'conds' key containing items with len >= 3
valid_input = {
    'conds': [
        [1, 2, 3],
        [4, 5, 6, 7],
        ['a', 'b', 'c']
    ],
    'other_key': 'some_value'
}

ARGS = {
    'x': valid_input
}
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
count_params Ran
ran on a synthesized input
salesforce/decaNLP/util.py
code served (permissive licence) · get_code("d7b2b4bb89a82c89")
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 torch.nn as nn

torch.manual_seed(0)

# The function count_params takes 'params' and iterates through them,
# calling .size() on each parameter to get its shape, then multiplying
# dimensions together to count total parameters.
# So params should be a list (or iterable) of torch tensors or nn.Parameters

# Create a list of tensors with distinct sizes to test the function
param1 = torch.randn(4, 3)  # 12 parameters
param2 = torch.randn(5, 2)  # 10 parameters
param3 = torch.randn(6,)    # 6 parameters

params = [param1, param2, param3]

ARGS = {
    "params": params
}
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 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
get_clip Ran
ran on a synthesized input
salesforce/decaNLP/validate.py
code served (permissive licence) · get_code("8ee61593cb444310")
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

class ValIter:
    def __init__(self, extra):
        self.extra = extra

ARGS = {
    "val_iter": ValIter(extra=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 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
get_context_question Ran
ran on a synthesized input
salesforce/decaNLP/util.py
code served (permissive licence) · get_code("3544f48b886804d2")
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)

class Example:
    def __init__(self):
        self.context_special = "context_special_token"
        self.context = "This is the context text"
        self.question_special = "question_special_token"
        self.question = "What is the answer?"

ex = Example()
context = "This is the context text"
question = "What is the answer?"
field = "text_field"

ARGS = {
    "ex": ex,
    "context": context,
    "question": question,
    "field": field
}
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
get_trainable_params Ran
ran on a synthesized input
eric-erki/decaNLP/util.py
pointer only (licence: BSD-3-Clause) · get_code("9289fb8937bd5a0c")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, eric-erki/decaNLP, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

matmul Ran
ran; our contract draft was wrong, not the code
eric-erki/decaNLP/models/common.py
pointer only (licence: BSD-3-Clause) · get_code("a211fd7add3a8e3e")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, eric-erki/decaNLP, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

pad_to_match Ran
ran on a synthesized input
salesforce/decaNLP/models/common.py
code served (permissive licence) · get_code("2e17682f4f73dc67")
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 3, 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

torch.manual_seed(0)

# Create two tensors with different sequence lengths
# x: batch_size=2, seq_len=3
# y: batch_size=2, seq_len=5
x = torch.randn(2, 3)
y = torch.randn(2, 5)

ARGS = {
    "x": x,
    "y": y,
}
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 3, 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
pad_to_match Ran
ran on a synthesized input
eric-erki/decaNLP/models/common.py
pointer only (licence: BSD-3-Clause) · get_code("b828330541293092")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, eric-erki/decaNLP, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

positional_encodings_like Ran
ran on a synthesized input
salesforce/decaNLP/models/common.py
code served (permissive licence) · get_code("a38860d700f83b59")
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 Tensor, ndim 2, dtype family float, shape (3, 4) as a later run recorded it. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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
from torch.autograd import Variable

torch.manual_seed(0)

# Create a small tensor with shape (batch=2, sequence_length=3, channels=4)
# The function expects x to have:
# - x.size(1) for sequence length
# - x.size(-1) for number of channels
# - x.size(2) for what appears to be channels again (based on usage)
x = torch.randn(2, 3, 4)

# Optional parameter t - if provided, should be position indices
t = torch.arange(0.0, 3)

ARGS = {
    "x": x,
    "t": t,
}
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 Tensor, ndim 2, dtype family float, shape (3, 4) — shape (3, 4) 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
positional_encodings_like Ran
ran on a synthesized input
eric-erki/decaNLP/models/common.py
pointer only (licence: BSD-3-Clause) · get_code("27fe94dcf3228e77")
How it ran

The body of this sample is withheld: BSD-3-Clause, as GitHub detects it on this paper's repository, eric-erki/decaNLP, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

score Ran
ran on a synthesized input
salesforce/decaNLP/metrics.py
code served (permissive licence) · get_code("e81a8138a9405edf")
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 ndarray, ndim 1, dtype family int, shape (4,) 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 string
import numpy as np

# Construct valid arguments for the score function
# The function takes two parameters: answer (str) and gold (list of str)

answer = "The quick brown fox jumps over the lazy dog"
gold = ["quick brown fox", "lazy dog"]

ARGS = {
    "answer": answer,
    "gold": gold
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family int, shape (4,) — shape (4,) 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
set_seed Ran
ran on a synthesized input
salesforce/decaNLP/util.py
code served (permissive licence) · get_code("d871f8c96e99be20")
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 device. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 random
import numpy as np

class Args:
    def __init__(self):
        self.seed = 42
        self.devices = [-1]  # CPU only

ARGS = {
    'args': Args(),
    'rank': None
}
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 device — a device output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a device 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 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, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 later) and 11 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
matmul 1 other paper
1 later
2312.01575 (2023-12)
A Challenging Multimodal Video Summary: Simultaneously Extracting and Generating Keyframe-Caption Pairs from Video
score 1 other paper
1 later
1812.06162 (2018-12)
An Empirical Model of Large-Batch Training

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-21.

Repositories linked to this paper

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

Abstract

Deep learning has improved performance on many natural language processing (NLP) tasks individually. However, general NLP models cannot emerge within a paradigm that focuses on the particularities of a single metric, dataset, and task. We introduce the Natural Language Decathlon (decaNLP), a challenge that spans ten tasks: question answering, machine translation, summarization, natural language inference, sentiment analysis, semantic role labeling, zero-shot relation extraction, goal-oriented dialogue, semantic parsing, and commonsense pronoun resolution. We cast all tasks as question answering over a context. Furthermore, we present a new Multitask Question Answering Network (MQAN) jointly learns all tasks in decaNLP without any task-specific modules or parameters in the multitask setting. MQAN shows improvements in transfer learning for machine translation and named entity recognition, domain adaptation for sentiment analysis and natural language inference, and zero-shot capabilities for text classification. We demonstrate that the MQAN's multi-pointer-generator decoder is key to this success and performance further improves with an anti-curriculum training strategy. Though designed for decaNLP, MQAN also achieves state of the art results on the WikiSQL semantic parsing task in the single-task setting. We also release code for procuring and processing data, training and evaluating models, and reproducing all experiments for decaNLP.

For agents

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

get_harvested_code_for_paper("1806.08730")
get_code_for_paper("1806.08730")
have("1806.08730")

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

curl https://syntology.ai/api/ran/1806.08730.json

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

[![Syntology run record](https://syntology.ai/api/ran/1806.08730.svg)](https://syntology.ai/paper/1806.08730)

Connect an agent — have() is free.