SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1904.04792 · 2019

Quizbowl: The Case for Incremental Question Answering

arXiv · PDF · Open in the Atlas

Code that ran

14 of 14 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 14 have no contract verdict; 0 failures 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/1904.04792.json · badge: /api/ran/1904.04792.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
h-j-han/automatic_explicitation community (classifier label)
classifier basis not recorded on the link
14 ran · 0 no recorded run
FunctionStatusWhere it lives
add_highlight_tokidx Ran
ran on a synthesized input
h-j-han/automatic_explicitation/show.py
code served (permissive licence) · get_code("ee20d0700e2f61f7")
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
import numpy as np
from typing import List

# Set seed for reproducibility
torch.manual_seed(0)

# Construct arguments for add_highlight_tokidx
src_tok = ['hello', 'world', 'this', 'is', 'a', 'test']
slne = [0, 2, 4]
color_name = 'GREEN'

ARGS = {
    'src_tok': src_tok,
    'slne': slne,
    'color_name': color_name
}
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 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
create_tokenizer Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/tokenizer.py
code served (permissive licence) · get_code("a3254eef985754fe")
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 type. 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 logging

logging.basicConfig(level=logging.INFO)

ARGS = {
    'lang': 'en',
    'name': 'whitespace'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
MeCab, sacremoses, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type type — a type output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a type 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
create_tokenizer4pairs Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/tokenizer.py
code served (permissive licence) · get_code("3c2ffcbb4d795cb4")
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 logging

logging.basicConfig(level=logging.INFO)

# Setup for create_tokenizer4pairs
# This function takes src_lang, tgt_lang, and pretokname as parameters
# and calls create_tokenizer for each language

# Valid language/tokenizer combinations based on the function logic:
# - lang='ja' returns MecabTokenizer (but requires MeCab library)
# - name='mbertcased' returns MultilingBertCasedBasicTokenizer
# - name='whitespace' returns WhitespaceTokenizer
# - name='moses' returns MosesTokenizer (but requires sacremoses library)

# To avoid external dependencies, we'll use 'whitespace' which has no external deps

ARGS = {
    'src_lang': 'en',
    'tgt_lang': 'fr',
    'pretokname': 'whitespace'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
MeCab, sacremoses, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
find_sent_idx Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/xqb/display_explicitation.py
code served (permissive licence) · get_code("d501e482699c498b")
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 numpy as np

# Analyze the function:
# - sents_seg: iterable of tuples (s, e) representing sentence start and end indices
# - start_idx: an integer index to search for within the segments
# - The function iterates through sents_seg looking for a segment where s <= start_idx <= e

# Create valid test data
sents_seg = [
    (0, 10),      # sentence 0: indices 0-10
    (11, 25),     # sentence 1: indices 11-25
    (26, 40),     # sentence 2: indices 26-40
]

# start_idx should be within one of the ranges
start_idx = 15  # falls within sentence 1 (11-25)

ARGS = {
    'sents_seg': sents_seg,
    'start_idx': start_idx,
}
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 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_token_char_indices Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/utils.py
code served (permissive licence) · get_code("9820556ecac3062a")
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 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

# Define ARGS for the get_token_char_indices function
# The function takes:
# - sent: a string (a sentence)
# - toks: a list of tokens (strings)

sent = "I ate apple."
toks = ["I", "ate", "apple", "."]

ARGS = {
    "sent": sent,
    "toks": toks
}
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 list, ndim 2, 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
get_unaligned_idx Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/utils.py
code served (permissive licence) · get_code("c1968e7cd9190083")
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 2, 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.

# Analyze the function to understand what it needs
# get_unaligned_idx takes:
# 1. align_words: List[tuple] - list of (i, j) pairs representing word alignments
# 2. src_tok: List[str] - source tokens (list of strings)
# 3. tgt_tok: List[str] - target tokens (list of strings)
#
# The function finds which token indices are NOT aligned between source and target.
# It iterates through align_words tuples, builds sets of aligned indices,
# then returns unaligned indices.

# Create test data
align_words = [(0, 0), (1, 1), (2, 2)]  # Source indices 0,1,2 aligned with target 0,1,2
src_tok = ["hello", "world", "there", "friend"]  # 4 source tokens (indices 0-3)
tgt_tok = ["bonjour", "monde", "la", "ami"]  # 4 target tokens (indices 0-3)
# Expected: src_unaligned = [3], tgt_unaligned = [3]

ARGS = {
    "align_words": align_words,
    "src_tok": src_tok,
    "tgt_tok": tgt_tok,
}
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 tuple, ndim 2, dtype family int — 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
group_ner Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/ner.py
code served (permissive licence) · get_code("99b73eb7e1a4905d")
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)

# Analyzing the function:
# - nes: list of dicts with 'entity' and 'index' keys
# - token: list of lists (word_list for each word)
# - printwarn: boolean flag (not used in function body)

# The function processes NER (Named Entity Recognition) outputs
# Each ne dict has:
#   - 'entity': string like 'B-PER', 'I-PER', 'O' (but 'O' causes assertion)
#   - 'index': integer (1-indexed based on usage)

# token is used to map token indices to word indices
# Each word can have multiple sub-tokens (subword tokenization)

nes = [
    {'entity': 'B-PER', 'index': 1},
    {'entity': 'I-PER', 'index': 2},
    {'entity': 'B-LOC', 'index': 3},
    {'entity': 'I-LOC', 'index': 4},
    {'entity': 'B-ORG', 'index': 5},
]

# token: each word can be split into multiple subword tokens
# For example, word 0 has 1 subtoken, word 1 has 2 subtokens, etc.
token = [
    ['token0'],           # word 0: 1 subtoken (indices 1)
    ['token1', 'token1_'],  # word 1: 2 subtokens (indices 2-3)
    ['token2'],           # word 2: 1 subtoken (index 4)
    ['token3', 'token3_'],  # word 3: 2 subtokens (indices 5-6)
    ['token4'],           # word 4: 1 subtoken (index 7)
]

ARGS = {
    'nes': nes,
    'token': token,
    'printwarn': False,
}
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
insert_token_make_batchsents Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/entitylink.py
code served (permissive licence) · get_code("fe80f19e05d4e058")
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 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

# Analyze the function:
# - tgt_tok: iterated over with enumerate, so needs to be a sequence of tokens (strings)
# - tlnetup: iterated over to unpack (s, e, t), so needs to be a list of tuples with 3 elements each
#   where s and e are integers used for indexing comparison with i
#   t is unpacked but not used in the function body

# Construct valid arguments
tgt_tok = ['hello', 'world', 'this', 'is', 'a', 'test']

# tlnetup: list of tuples (start_idx, end_idx, unused_value)
# Let's create a few examples with different start/end positions
tlnetup = [
    (1, 3, 'label1'),
    (0, 5, 'label2'),
    (2, 4, 'label3'),
]

ARGS = {
    'tgt_tok': tgt_tok,
    'tlnetup': tlnetup,
}
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 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
plot3 Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/plots.py
code served (permissive licence) · get_code("cca6327af9af1e77")
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 Figure. 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 matplotlib.pyplot as plt
from matplotlib.figure import Figure
import numpy as np

torch.manual_seed(0)

# Define a minimal plot function that matches the expected signature
def dummy_plot_func(ax, lcompare, data, **kwargs):
    """Minimal plotting function that accepts the required arguments"""
    ax.plot([1, 2, 3], [1, 2, 3])
    ax.set_title(f"Plot with {len(lcompare)} items")

# Create minimal data structure
data = {
    "values": torch.randn(2, 4),
    "labels": ["a", "b"]
}

# Create comparison lists - two lists of items to compare
llcompare = [
    ["item1_a", "item1_b"],
    ["item2_a", "item2_b"],
    ["item3_a", "item3_b"]
]

# Create args for each subplot
largs = [
    {"color": "red"},
    {"color": "blue"},
    {"color": "green"}
]

# Additional kwargs to pass through
kwargs_extra = {"linewidth": 2}

ARGS = {
    "plot_func": dummy_plot_func,
    "llcompare": llcompare,
    "data": data,
    "largs": largs,
    "kwargs": kwargs_extra
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Figure — a Figure output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Figure 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
plot6 Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/plots.py
code served (permissive licence) · get_code("1b161f90a980b964")
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 module. 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 matplotlib.pyplot as plt
import numpy as np

torch.manual_seed(0)

# Define a simple plot function that will be passed as plot_func
def simple_plot_func(ax, lcompare, data, *args, **kwargs):
    """
    A minimal plot function that accepts ax, lcompare, and data.
    This function should work with the plot6 signature.
    """
    x = np.linspace(0, 10, 100)
    y = np.sin(x)
    ax.plot(x, y, label=str(lcompare))
    ax.set_title(f"Plot for {lcompare}")
    ax.legend()

# Create llcompare: a list of 6 comparisons (since we have 2x3 grid)
llcompare = [
    "compare_1",
    "compare_2",
    "compare_3",
    "compare_4",
    "compare_5",
    "compare_6"
]

# Create data: can be any object, we'll use a simple dict
data = {
    "values": torch.randn(2, 4),
    "labels": ["label1", "label2"]
}

# Optional args and kwargs for the plot function
args_list = []
kwargs_dict = {
    "color": "blue",
    "linewidth": 2.0
}

ARGS = {
    "plot_func": simple_plot_func,
    "llcompare": llcompare,
    "data": data,
    "args": args_list,
    "kwargs": kwargs_dict
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type module — a module output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a module 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
plot8 Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/plots.py
code served (permissive licence) · get_code("5983071885cc410d")
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 module. 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 matplotlib.pyplot as plt
from matplotlib.axes import Axes

torch.manual_seed(0)

# Define a simple plot function that will be called for each subplot
def simple_plot_func(ax, lcompare, data, *args, **kwargs):
    """Simple stand-in plot function that takes an axes object and data."""
    ax.plot([1, 2, 3], [1, 2, 3])
    ax.set_title(f"Plot {lcompare}")

# Create sample data - a small tensor
sample_data = torch.randn(2, 4)

# Create a list of compare items (one for each subplot, 8 items total for 2x4 grid)
llcompare = [
    ["compare_1a", "compare_1b"],
    ["compare_2a", "compare_2b"],
    ["compare_3a", "compare_3b"],
    ["compare_4a", "compare_4b"],
    ["compare_5a", "compare_5b"],
    ["compare_6a", "compare_6b"],
    ["compare_7a", "compare_7b"],
    ["compare_8a", "compare_8b"],
]

# Define ARGS dictionary with all parameters for plot8
ARGS = {
    "plot_func": simple_plot_func,
    "llcompare": llcompare,
    "data": sample_data,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type module — a module output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a module 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
print_span Ran
ran on a synthesized input
h-j-han/automatic_explicitation/show.py
code served (permissive licence) · get_code("8c9acd0ab7ead2bd")
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
import numpy as np
from typing import List

# Define ARGS dict with valid arguments for print_span function
ARGS = {
    'lsrcbrack': [(0, 2), (4, 5)],  # List of tuples representing index ranges
    'srctok_orig': ['hello', 'world', 'this', 'is', 'a', 'test'],  # List of string tokens
    'color': 'GREEN'  # Color name as string
}
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 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
remove_duplicate_tuples Ran
ran on a synthesized input
h-j-han/automatic_explicitation/autoexpl/utils.py
code served (permissive licence) · get_code("8d1d41b003e7a86e")
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

# Analyze the function:
# - It takes a list `lst` as input
# - It iterates through elements (called `tup`)
# - For each element, it accesses `tup[:1]` (slicing the first element)
# - It checks if `tup[:1]` is in a set
# - It appends `tup` to result and adds `tup[:1]` to seen set
# The function appears to deduplicate based on the first element of each tuple

# Create a list of tuples where we can deduplicate by first element
# Use tuples with distinct first elements and varying second elements
lst = [
    (1, 'a'),
    (2, 'b'),
    (1, 'c'),  # duplicate first element, should be filtered
    (3, 'd'),
    (2, 'e'),  # duplicate first element, should be filtered
    (4, 'f'),
]

ARGS = {
    'lst': lst
}
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
tup_range_to_idx Ran
ran on a synthesized input
h-j-han/automatic_explicitation/show.py
code served (permissive licence) · get_code("f84f0926a1f7a503")
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

def tup_range_to_idx(tupleidxrange):
    slne = []
    for t in tupleidxrange:
        s = t[0]
        e = t[1]
        for i in range(s, e + 1):
            slne.append(i)
    return slne

torch.manual_seed(0)

ARGS = {
    'tupleidxrange': [(0, 2), (5, 7), (10, 12)]
}
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 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

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 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.

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

Scholastic trivia competitions test knowledge and intelligence through mastery of question answering. Modern question answering benchmarks are one variant of the Turing test. Specifically, answering a set of questions as well as a human is a minimum bar towards demonstrating human-like intelligence. This paper makes the case that the format of one competition -- where participants can answer in the middle of hearing a question (incremental) -- better differentiates the skill between (human or machine) players. Additionally, merging a sequential decision-making sub-task with question answering (QA) provides a good setting for research in model calibration and opponent modeling. Thus, embedded in this task are three machine learning challenges: (1) factoid QA over thousands of Wikipedia-like answers, (2) calibration of the QA model's confidence scores, and (3) sequential decision-making that incorporates knowledge of the QA model, its calibration, and what the opponent may do. We make two contributions: (1) collecting and curating a large factoid QA dataset and an accompanying gameplay dataset, and (2) developing a model that addresses these three machine learning challenges. In addition to offline evaluation, we pitted our model against some of the most accomplished trivia players in the world in a series of exhibition matches spanning several years. Throughout this paper, we show that collaborations with the vibrant trivia community have contributed to the quality of our dataset, spawned new research directions, and doubled as an exciting way to engage the public with research in machine learning and natural language processing.

For agents

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

get_harvested_code_for_paper("1904.04792")
get_code_for_paper("1904.04792")
have("1904.04792")

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

curl https://syntology.ai/api/ran/1904.04792.json

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

[![Syntology run record](https://syntology.ai/api/ran/1904.04792.svg)](https://syntology.ai/paper/1904.04792)

Connect an agent — have() is free.