SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1901.04085 · 2019

Passage Re-ranking with BERT

arXiv · PDF · Open in the Atlas

Code that ran

11 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 8 have no contract verdict; 3 failures of our instrument; 2 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/1901.04085.json · badge: /api/ran/1901.04085.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
nyu-dl/dl4marco-bert canonical (classifier label)
classifier basis: census
11 ran · 2 no recorded run
FunctionStatusWhere it lives
NDCG Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/metrics.py
code served (permissive licence) · get_code("1e33e4465ad9760e")
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 float64, ndim 0, 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 numpy as np

# Construct arguments for NDCG function
# The function expects:
# - gt: a list (or list-like) of ground truth items that supports .index() method
# - pred: a list (or list-like) of predicted items
# - use_graded_scores: boolean flag

# Create ground truth list with some items
gt = ['item_a', 'item_b', 'item_c', 'item_d']

# Create prediction list with some overlap with gt
pred = ['item_b', 'item_x', 'item_a', 'item_y', 'item_c']

# Boolean flag for graded scores
use_graded_scores = False

ARGS = {
    'gt': gt,
    'pred': pred,
    'use_graded_scores': use_graded_scores
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
average_precision Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/metrics.py
code served (permissive licence) · get_code("45e8829b47e5a1b7")
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)

# Based on the function signature and body:
# - gt: a set of ground-truth elements
# - pred: a list of predicted elements
# Both contain comparable elements (e.g., integers or strings)

# Create ground truth set with some example elements
gt = {1, 3, 5, 7}

# Create prediction list with some overlap and some new elements
# Order matters for pred, and duplicates should not be counted twice
pred = [1, 2, 3, 4, 5, 6]

ARGS = {
    "gt": gt,
    "pred": pred
}
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
convert_to_bert_input Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/tokenization.py
code served (permissive licence) · get_code("895873eb21616a05")
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 int; 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 MockTokenizer:
    def __init__(self):
        self.vocab = {
            '[CLS]': 101,
            '[SEP]': 102,
            'hello': 7592,
            'world': 2088,
            'this': 2023,
            'is': 2003,
            'a': 1037,
            'test': 3231,
        }
    
    def tokenize(self, text):
        words = text.lower().split()
        tokens = []
        for word in words:
            if word in self.vocab:
                tokens.append(word)
            else:
                tokens.append('[UNK]')
        return tokens
    
    def convert_tokens_to_ids(self, tokens):
        ids = []
        for token in tokens:
            if token in self.vocab:
                ids.append(self.vocab[token])
            else:
                ids.append(100)
        return ids

text = "hello world this is a test"
max_seq_length = 10
tokenizer = MockTokenizer()
add_cls = True

ARGS = {
    'text': text,
    'max_seq_length': max_seq_length,
    'tokenizer': tokenizer,
    'add_cls': add_cls,
}
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 int — 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
convert_to_unicode Ran
ran; our contract draft was wrong, not the code
nyu-dl/dl4marco-bert/tokenization.py
pointer only (licence: BSD-3-Clause); the stored copy get_code serves, from huawei-noah/Pretrained-Language-Model, is pointer only (licence: NONE) · get_code("1923fc05163d207d")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-3-Clause), but get_code serves the stored copy, first harvested from huawei-noah/Pretrained-Language-Model, and GitHub detected no licence in that repository when it was checked. 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.

get_activation Ran
ran; our contract draft was wrong, not the code
nyu-dl/dl4marco-bert/modeling.py
code served (permissive licence) · get_code("5681f7dd9b6e3679")
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) 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 tensorflow as tf
import six

# Test cases for get_activation function
# The function takes activation_string as parameter and returns either None or a callable

ARGS = {
    "activation_string": "relu"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
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
input_fn_builder Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/run_msmarco.py
code served (permissive licence) · get_code("e010030ee57fe701")
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 tensorflow as tf
import tempfile
import os

# Create a minimal TFRecord file for testing
def create_test_tfrecord(file_path):
    """Create a small test TFRecord file with sample data."""
    writer = tf.io.TFRecordWriter(file_path)
    
    # Create a few sample records
    for i in range(3):
        feature = {
            'query_ids': tf.train.Feature(int64_list=tf.train.Int64List(value=[1, 2])),
            'doc_ids': tf.train.Feature(int64_list=tf.train.Int64List(value=[3, 4, 5])),
            'label': tf.train.Feature(int64_list=tf.train.Int64List(value=[1])),
        }
        example = tf.train.Example(features=tf.train.Features(feature=feature))
        writer.write(example.SerializeToString())
    
    writer.close()

# Create temporary TFRecord file
temp_dir = tempfile.mkdtemp()
dataset_path = os.path.join(temp_dir, "test_data.tfrecord")
create_test_tfrecord(dataset_path)

ARGS = {
    'dataset_path': dataset_path,
    'seq_length': 8,
    'is_training': False,
    'max_eval_examples': 2,
}
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 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
input_fn_builder Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/run_treccar.py
code served (permissive licence) · get_code("9fbf56276f399498")
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 tensorflow as tf
import tempfile
import os

# Create a minimal TFRecord file for testing
def create_dummy_tfrecord(file_path):
    """Create a small TFRecord file with sample data."""
    writer = tf.io.TFRecordWriter(file_path)
    
    # Create a sample record
    feature = {
        'query_ids': tf.train.Feature(int64_list=tf.train.Int64List(value=[101, 102, 103])),
        'doc_ids': tf.train.Feature(int64_list=tf.train.Int64List(value=[104, 105])),
        'label': tf.train.Feature(int64_list=tf.train.Int64List(value=[1])),
        'len_gt_titles': tf.train.Feature(int64_list=tf.train.Int64List(value=[2])),
    }
    
    example = tf.train.Example(features=tf.train.Features(feature=feature))
    writer.write(example.SerializeToString())
    writer.close()

# Create temporary directory and TFRecord file
temp_dir = tempfile.mkdtemp()
dataset_path = os.path.join(temp_dir, 'dummy.tfrecord')
create_dummy_tfrecord(dataset_path)

ARGS = {
    'dataset_path': dataset_path,
    'seq_length': 16,
    'is_training': True,
    'max_eval_examples': None,
}
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 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
load_qrels Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/convert_treccar_to_tfrecord.py
code served (permissive licence) · get_code("9eb250ad7f25c79e")
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 defaultdict. 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 collections
import tempfile
import os

# Create a temporary file with qrels data
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_path = temp_file.name

# Write sample qrels data in the expected format
# Format: topic query_id doc_title relevance
qrels_data = """topic1 0 doc1 1
topic1 0 doc2 2
topic2 0 doc3 1
topic1 0 doc4 0
topic3 0 doc5 1
"""

temp_file.write(qrels_data)
temp_file.close()

ARGS = {
    'path': temp_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 defaultdict — a defaultdict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a defaultdict 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_run Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/convert_treccar_to_tfrecord.py
code served (permissive licence) · get_code("86c70a8984ca78b7")
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.

import tempfile
import collections
import os

# Create a temporary file with run data in the expected format
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.txt')
temp_path = temp_file.name

# Write sample run data in the expected format:
# topic Q0 doc_id rank score run_id
run_data = [
    "topic1 Q0 doc001 1 10.5 run1",
    "topic1 Q0 doc002 2 9.3 run1",
    "topic1 Q0 doc003 3 8.1 run1",
    "topic2 Q0 doc004 1 11.2 run1",
    "topic2 Q0 doc005 2 10.0 run1",
    "topic3 Q0 doc006 1 12.5 run1",
]

for line in run_data:
    temp_file.write(line + '\n')

temp_file.close()

ARGS = {
    'path': temp_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 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
metrics Ran
ran on a synthesized input
nyu-dl/dl4marco-bert/metrics.py
code served (permissive licence) · get_code("54573d73f899d22f")
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 float, shape (5,) 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

# Analyze the function signature:
# metrics(gt, pred, metrics_map)
# 
# From the body:
# - gt: must support set operations (&), len(), and iteration
# - pred: must be a list, supports indexing and iteration
# - metrics_map: a list that supports .index() method

# gt is used as a set in operations like: gt & set(pred[:len(gt)])
# and checking membership: if item in gt
# So gt should be a set of hashable items

# pred is a list of items that should match items in gt
# metrics_map is a list of metric names

# Create simple test data
gt = {'item_1', 'item_2', 'item_3', 'item_5'}  # ground truth set
pred = ['item_1', 'item_4', 'item_2', 'item_5', 'item_6']  # predicted list
metrics_map = ['MAP', 'RPrec', 'MRR', 'MRR@10', 'NDCG']  # list of metric names

ARGS = {
    'gt': gt,
    'pred': pred,
    'metrics_map': metrics_map
}
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 float, shape (5,) — shape (5,) 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
printable_text Ran
ran; our contract draft was wrong, not the code
nyu-dl/dl4marco-bert/tokenization.py
code served (permissive licence) · get_code("0e5615f8994003cf")
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 six

# The function takes a single parameter 'text' which can be:
# - str (works in both Python 2 and 3)
# - bytes (in Python 3, decoded to utf-8)
# - unicode (in Python 2, encoded to utf-8)

# Since we're running on Python 3, we'll test with a string
text_input = "Hello, World!"

ARGS = {
    "text": text_input
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
six — 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
gelu Not yet run
unverified (harvested, no recorded run)
nyu-dl/dl4marco-bert/modeling.py
code served (permissive licence) · get_code("ecab128238ebf253")
How it ran

no invocation recorded.

get_assignment_map_from_checkpoint Not yet run
unverified (harvested, no recorded run)
nyu-dl/dl4marco-bert/modeling.py
pointer only (licence: BSD-3-Clause); the stored copy get_code serves, from toggle1995/RIS-DMMI, is pointer only (licence: NONE) · get_code("50958618b65e514e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-3-Clause), but get_code serves the stored copy, first harvested from toggle1995/RIS-DMMI, and GitHub detected no licence in that repository when it was checked. 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. No invocation recorded.

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, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 5 the same month, 5 later) and 8 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
convert_to_unicode 36 other papers
2 dated earlier, 1 the same month, 33 later
1609.08144 (2016-09)
Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.08746 (2019-01)
BioBERT: a pre-trained biomedical language representation model for biomedical text mining
and 33 more not named here
printable_text 33 other papers
2 dated earlier, 1 the same month, 30 later
1609.08144 (2016-09)
Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.08746 (2019-01)
BioBERT: a pre-trained biomedical language representation model for biomedical text mining
and 30 more not named here
get_activation 32 other papers
2 dated earlier, 1 the same month, 29 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
1901.08746 (2019-01)
BioBERT: a pre-trained biomedical language representation model for biomedical text mining
and 29 more not named here
get_assignment_map_from_checkpoint 26 other papers
3 dated earlier, 1 the same month, 22 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
gelu 16 other papers
3 dated earlier, 1 the same month, 12 later · 2 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
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
and 13 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 2026-11-21.

Repositories linked to this paper

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

Abstract

Recently, neural models pretrained on a language modeling task, such as ELMo (Peters et al., 2017), OpenAI GPT (Radford et al., 2018), and BERT (Devlin et al., 2018), have achieved impressive results on various natural language processing tasks such as question-answering and natural language inference. In this paper, we describe a simple re-implementation of BERT for query-based passage re-ranking. Our system is the state of the art on the TREC-CAR dataset and the top entry in the leaderboard of the MS MARCO passage retrieval task, outperforming the previous state of the art by 27% (relative) in MRR@10. The code to reproduce our results is available at https://github.com/nyu-dl/dl4marco-bert

For agents

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

get_harvested_code_for_paper("1901.04085")
get_code_for_paper("1901.04085")
have("1901.04085")

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

curl https://syntology.ai/api/ran/1901.04085.json

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

[![Syntology run record](https://syntology.ai/api/ran/1901.04085.svg)](https://syntology.ai/paper/1901.04085)

Connect an agent — have() is free.