SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1911.00484 · 2019

Select, Answer and Explain: Interpretable Multi-hop Reading Comprehension over Multiple Documents

arXiv · PDF · Open in the Atlas

Code that ran

12 of 15 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 11 have no contract verdict; 1 failure of our instrument; 3 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/1911.00484.json · badge: /api/ran/1911.00484.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
jd-ai-research-silicon-valley/sae community (classifier label)
classifier basis not recorded on the link
12 ran · 3 no recorded run
FunctionStatusWhere it lives
batchify Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/dataset.py
code served (permissive licence) · get_code("f02424fc99b0d00d")
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. 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)

# Create a single example with all required fields
# ex[0]: input_ids tensor (num_docs, seq_len)
# ex[1]: segment_ids tensor (num_docs, seq_len)
# ex[2]: input_mask tensor (num_docs, seq_len)
# ex[3]: labels tensor (num_pairs,)
# ex[4]: triplet tensor (num_pairs, 2)
# ex[5]: true_labels tensor (scalar or 1D)

example1 = (
    torch.randint(0, 100, (3, 5)).long(),  # input_ids: 3 docs, 5 seq_len
    torch.randint(0, 2, (3, 5)).long(),    # segment_ids: 3 docs, 5 seq_len
    torch.randint(0, 2, (3, 5)).long(),    # input_mask: 3 docs, 5 seq_len
    torch.tensor([0.5, 0.3]).float(),      # labels: 2 pairs
    torch.randint(0, 3, (2, 2)).long(),    # triplet: 2 pairs, 2 indices
    torch.tensor(1)                        # true_labels: scalar
)

example2 = (
    torch.randint(0, 100, (2, 5)).long(),  # input_ids: 2 docs, 5 seq_len
    torch.randint(0, 2, (2, 5)).long(),    # segment_ids: 2 docs, 5 seq_len
    torch.randint(0, 2, (2, 5)).long(),    # input_mask: 2 docs, 5 seq_len
    torch.tensor([0.7]).float(),           # labels: 1 pair
    torch.randint(0, 3, (1, 2)).long(),    # triplet: 1 pair, 2 indices
    torch.tensor(0)                        # true_labels: scalar
)

batch = [example1, example2]

ARGS = {
    "batch": batch
}
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 — a tuple 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
clear_punc Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/combine_pred.py
code served (permissive licence) · get_code("77c73d18a16d7d3b")
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

def clear_punc(text):
    puncs = [',', '-', "'", '.', '?', '!']
    new_text = ''
    for i in range(0, len(text)):
        if text[i] == ' ' and (text[i - 1] in puncs or text[i + 1] in puncs):
            continue
        else:
            new_text += text[i]
    return new_text

ARGS = {
    'text': "Hello, world - it's a beautiful day!"
}
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
context_extractor Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/NERExtractorTest.py
code served (permissive licence) · get_code("68fd38afe5e40eea")
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 Series. 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 pandas as pd

# Construct a valid doc_list argument
# Each doc is a tuple/list where:
# - doc[0] is a context name (string)
# - doc[1] is a list of sentences (the length is what matters)

doc_list = [
    ("document_1", ["sentence 1", "sentence 2", "sentence 3"]),
    ("document_2", ["sentence 1", "sentence 2"]),
    ("document_3", ["sentence 1", "sentence 2", "sentence 3", "sentence 4"])
]

ARGS = {
    "doc_list": doc_list
}
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 Series — a Series output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Series output the recorded ndim 1 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
find_facts Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/prepro.py
code served (permissive licence) · get_code("1ac491185ac31260")
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 OrderedDict. 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 collections import OrderedDict

def find_facts(ex):
    doc_dict = OrderedDict()
    for i, doc in enumerate(ex['context']):
        doc_dict[doc[0]] = [i, 0]
    for i, id_sent in enumerate(ex['supporting_facts']):
        if doc_dict.get(id_sent[0]) is not None:
            doc_dict[id_sent[0]][1] = 1
            if ex['answer'] not in ['yes', 'no']:
                index = doc_dict[id_sent[0]][0]
                if ex['answer'] in ''.join(ex['context'][index][1]):
                    doc_dict[id_sent[0]][1] = 2
    return doc_dict


ARGS = {
    'ex': {
        'context': [
            ('Document A', ['The cat sat on the mat', 'It was very comfortable']),
            ('Document B', ['Dogs are loyal', 'They love their owners']),
            ('Document C', ['The answer is here', 'This contains the solution']),
        ],
        'supporting_facts': [
            ['Document A', 0],
            ['Document C', 1],
        ],
        'answer': 'the solution',
    }
}
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 OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict 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
find_test_facts Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/prepro.py
code served (permissive licence) · get_code("e9aa7eccda0f33fd")
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 OrderedDict. 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 collections import OrderedDict

# Construct valid arguments for find_test_facts
# The function expects 'ex' with keys: 'context', 'hop1_query', 'hop2_query'

ARGS = {
    'ex': {
        'context': [
            ('query_one', 'some text'),
            ('query_two', 'more text'),
            ('other_doc', 'different text'),
            ('another_doc', 'yet more text'),
        ],
        'hop1_query': 'query_one',
        'hop2_query': 'query_two',
    }
}
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 OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict 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
gen_adj_matrix Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/hotpotqa_loader.py
code served (permissive licence) · get_code("51aead194c7e350e")
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 4, dtype family float, shape (2, 3, 4, 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 numpy as np
import torch
from scipy.sparse import coo_matrix

torch.manual_seed(0)

class Feature:
    def __init__(self, wd_edges, ad_edges, ques_edges):
        self.wd_edges = wd_edges
        self.ad_edges = ad_edges
        self.ques_edges = ques_edges

# Create sample features with edge lists
feature1 = Feature(
    wd_edges=[(0, 1), (1, 2)],
    ad_edges=[(0, 2), (2, 3)],
    ques_edges=[(1, 3)]
)

feature2 = Feature(
    wd_edges=[(0, 1)],
    ad_edges=[],
    ques_edges=[(0, 2), (1, 3)]
)

train_features = [feature1, feature2]
max_sent_num = 4

ARGS = {
    'train_features': train_features,
    'max_sent_num': max_sent_num,
    'adj_norm': False,
    'wdedge': True,
    'adedge': True,
    'quesedge': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 4, dtype family float, shape (2, 3, 4, 4) — shape (2, 3, 4, 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
get_features Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/prepro.py
code served (permissive licence) · get_code("4196e08cc75d063a")
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. 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 tempfile
import os

class MockTokenizer:
    def tokenize(self, text):
        # Simple whitespace tokenization
        return text.lower().split()
    
    def convert_tokens_to_ids(self, tokens):
        # Simple mapping: each token gets a unique id based on hash
        return [hash(token) % 1000 for token in tokens]

# Create a temporary JSON file with valid data
temp_data = [
    {
        "question": "What is the capital?",
        "context": [
            [
                "Title1",
                [
                    "This is the first sentence.",
                    "This is the second sentence."
                ]
            ],
            [
                "Title2",
                [
                    "Another document here.",
                    "With multiple sentences."
                ]
            ]
        ],
        "supporting_facts": [
            ["Title1", 0],
            ["Title2", 1]
        ],
        "answer": "yes"
    },
    {
        "question": "Where is it located?",
        "context": [
            [
                "Title3",
                [
                    "Location information goes here.",
                    "More details about the place."
                ]
            ]
        ],
        "supporting_facts": [
            ["Title3", 0]
        ],
        "answer": "Paris"
    }
]

# Create temporary file
temp_file = tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False)
json.dump(temp_data, temp_file)
temp_file.close()

ARGS = {
    "data_path": temp_file.name,
    "tokenizer": MockTokenizer(),
    "use_mini": False,
    "max_seq_len": 512
}
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 — a list 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_sp_pred Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/hotpotqa_utils_joint.py
code served (permissive licence) · get_code("b41fccaf4208ef3c")
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 2, 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

class MockDataItem:
    def __init__(self, doc_title, sent_id):
        self.doc_title = doc_title
        self.sent_id = sent_id

# Create mock data items with required attributes
data_item_1 = MockDataItem(doc_title=["Document A"], sent_id=0)
data_item_2 = MockDataItem(doc_title=["Document B"], sent_id=1)
data_item_3 = MockDataItem(doc_title=["Document C"], sent_id=2)

# Create data list
data = [data_item_1, data_item_2, data_item_3]

# Create pred_sp_idx - indices that select from data
# Mix indices that are within bounds and potentially out of bounds
pred_sp_idx = [0, 2, 1]

ARGS = {
    "pred_sp_idx": pred_sp_idx,
    "data": data
}
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 2, 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
loadJSonAsDataFrame Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/Core/IOUtils.py
code served (permissive licence) · get_code("37bfc0274ce68364")
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 DataFrame. 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 tempfile
import os
import pandas as pd

# Create a temporary JSON file with sample data
temp_dir = tempfile.mkdtemp()
json_file_path = os.path.join(temp_dir, "test_data.json")

# Create sample JSON data and write to file
sample_data = [
    {"id": 1, "name": "Alice", "value": 10.5},
    {"id": 2, "name": "Bob", "value": 20.3},
    {"id": 3, "name": "Charlie", "value": 15.7}
]

with open(json_file_path, 'w') as f:
    json.dump(sample_data, f)

ARGS = {
    "jsonFileName": json_file_path
}
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 DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
loadPickleAsDataFrame Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/Core/IOUtils.py
code served (permissive licence) · get_code("5a590a2f61ae3df9")
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 DataFrame. 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 pandas as pd
import tempfile
import os

# Create a temporary pickle file with a valid DataFrame
temp_dir = tempfile.mkdtemp()
test_df = pd.DataFrame({
    'col1': [1, 2, 3],
    'col2': ['a', 'b', 'c'],
    'col3': [1.5, 2.5, 3.5]
})
pickle_file_path = os.path.join(temp_dir, 'test.pkl')
test_df.to_pickle(pickle_file_path)

ARGS = {
    'pickle_file_name': pickle_file_path
}
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 DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
normalize_answer Ran
ran; our contract draft was wrong, not the code
jd-ai-research-silicon-valley/sae/hotpotqa_utils_joint.py
code served (permissive licence) · get_code("58c6794946fd1e26")
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 string
import re

def normalize_answer(s):

    def remove_articles(text):
        return re.sub('\\b(a|an|the)\\b', ' ', text)

    def white_space_fix(text):
        return ' '.join(text.split())

    def remove_punc(text):
        exclude = set(string.punctuation)
        return ''.join((ch for ch in text if ch not in exclude))

    def lower(text):
        return text.lower()
    return white_space_fix(remove_articles(remove_punc(lower(s))))


ARGS = {
    's': 'The Quick Brown Fox jumps over the lazy dog!'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
process_logit Ran
ran on a synthesized input
jd-ai-research-silicon-valley/sae/hotpotqa_utils_joint.py
code served (permissive licence) · get_code("b5bb8196fc9781a5")
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. 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)

# Create a small stand-in class for data items
class DataItem:
    def __init__(self, doc_title, sent_id):
        self.doc_title = doc_title
        self.sent_id = sent_id

# Create sample data structures
# predict_examples is a list of lists, where each inner list contains DataItem objects
predict_examples = [
    [
        DataItem(doc_title=["Document1"], sent_id=0),
        DataItem(doc_title=["Document1"], sent_id=1),
        DataItem(doc_title=["Document2"], sent_id=0),
        DataItem(doc_title=["Document2"], sent_id=1),
    ],
    [
        DataItem(doc_title=["Document3"], sent_id=0),
        DataItem(doc_title=["Document3"], sent_id=1),
        DataItem(doc_title=["Document4"], sent_id=0),
        DataItem(doc_title=["Document4"], sent_id=1),
    ]
]

# batch_index: tensor with indices into predict_examples
batch_index = torch.tensor([0, 1])

# batch_logits: tuple of (supporting_facts_logits, answer_type_logits)
# supporting_facts_logits: shape (batch_size=2, num_facts=4)
# answer_type_logits: shape (batch_size=2, num_types=3)
sp_logits = torch.randn(2, 4)
ans_type_logits = torch.randn(2, 3)
batch_logits = (sp_logits, ans_type_logits)

# predict_features: not used in the function body
predict_features = []

# max_answer_length: not used in the function body
max_answer_length = 100

ARGS = {
    "batch_index": batch_index,
    "batch_logits": batch_logits,
    "predict_features": predict_features,
    "predict_examples": predict_examples,
    "max_answer_length": max_answer_length,
}
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 tuple — a tuple 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
evaluate_multirc Not yet run
unverified (harvested, no recorded run)
jd-ai-research-silicon-valley/sae/hotpotqa_loader.py
code served (permissive licence) · get_code("17260cf580638054")
How it ran

no invocation recorded.

process_logit Not yet run
unverified (harvested, no recorded run)
jd-ai-research-silicon-valley/sae/hotpotqa_loader.py
code served (permissive licence) · get_code("c4023966c7eb0cda")
How it ran

no invocation recorded.

sample_accuracy_and_recall Not yet run
unverified (harvested, no recorded run)
jd-ai-research-silicon-valley/sae/docfilter.py
code served (permissive licence) · get_code("fcad9ec20082c970")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 1 later) and 14 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
normalize_answer 133 other papers
17 dated earlier, 3 the same month, 112 later, 1 undated · 1 only as an assembly brick
1508.06615 (2015-08)
Character-Aware Neural Language Models
1606.05250 (2016-06)
SQuAD: 100,000+ Questions for Machine Comprehension of Text
1608.07905 (2016-08)
Machine Comprehension Using Match-LSTM and Answer Pointer
and 130 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

Interpretable multi-hop reading comprehension (RC) over multiple documents is a challenging problem because it demands reasoning over multiple information sources and explaining the answer prediction by providing supporting evidences. In this paper, we propose an effective and interpretable Select, Answer and Explain (SAE) system to solve the multi-document RC problem. Our system first filters out answer-unrelated documents and thus reduce the amount of distraction information. This is achieved by a document classifier trained with a novel pairwise learning-to-rank loss. The selected answer-related documents are then input to a model to jointly predict the answer and supporting sentences. The model is optimized with a multi-task learning objective on both token level for answer prediction and sentence level for supporting sentences prediction, together with an attention-based interaction between these two tasks. Evaluated on HotpotQA, a challenging multi-hop RC data set, the proposed SAE system achieves top competitive performance in distractor setting compared to other existing systems on the leaderboard.

For agents

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

get_harvested_code_for_paper("1911.00484")
get_code_for_paper("1911.00484")
have("1911.00484")

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

curl https://syntology.ai/api/ran/1911.00484.json

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

[![Syntology run record](https://syntology.ai/api/ran/1911.00484.svg)](https://syntology.ai/paper/1911.00484)

Connect an agent — have() is free.