SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2207.00735 · NeurIPS · 2022

Can Language Models Make Fun? A Case Study in Chinese Comical Crosstalk

Benyou Wang, Xiangbo Wu, Xiaokang Liu, Jianquan Li, Prayag Tiwari, Qianqian Xie

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 9 have no contract verdict; 2 failures of our instrument; 2 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/2207.00735.json · badge: /api/ran/2207.00735.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
anonNo2/crosstalk-generation canonical (classifier label)
classifier basis not recorded on the link
11 ran · 2 no recorded run
FunctionStatusWhere it lives
annotate_heatmap Ran
ran; our contract draft was wrong, not the code
anonNo2/crosstalk-generation/src/beta_code/muti_heatmap_demo.py
code served (permissive licence) · get_code("fa46445bbe1d057f")
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 numpy as np
import matplotlib.pyplot as plt
import matplotlib.ticker

# Create sample data
data = np.random.randn(6, 6)

# Create a figure and axis
fig, ax = plt.subplots()

# Create an image on the axis
im = ax.imshow(data, cmap='viridis')

# Define arguments for annotate_heatmap
ARGS = {
    'im': im,
    'data': data,
    'valfmt': '{x:.2f}',
    'textcolors': ('black', 'white'),
    'threshold': None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — 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
calculate_bleu_score Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/beta_code/machine_metrics.py
code served (permissive licence) · get_code("e825114d21a034f8")
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 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.

from nltk.translate.bleu_score import corpus_bleu
from nltk.translate.bleu_score import SmoothingFunction

# Create sample data for the function
# references should be a list of reference sentences (each sentence is a list of tokens)
# candidates should be a list of candidate sentences (each sentence is a list of tokens)

references = [
    ["the", "cat", "sat", "on", "the", "mat"],
    ["a", "dog", "ran", "in", "the", "park"],
    ["birds", "fly", "high", "in", "the", "sky"]
]

candidates = [
    ["the", "cat", "sat", "on", "mat"],
    ["a", "dog", "runs", "in", "park"],
    ["birds", "fly", "high", "sky"]
]

ARGS = {
    "references": references,
    "candidates": candidates
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
calculate_gleu_score Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/beta_code/machine_metrics.py
code served (permissive licence) · get_code("3c24b85f295e82cd")
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)

# Analyze the function:
# - references: iterated over, each element i is iterated over to get j's
# - candidates: iterated over, each element i is iterated over to get j's
# - The function wraps references as [[[j for j in i]] for i in references]
# - The function wraps candidates as [[j for j in i] for i in candidates]
# - These are passed to corpus_gleu which expects:
#   - reference: list of reference sets (each reference set is a list of reference translations)
#   - hypothesis: list of hypothesis translations

# For GLEU score, we need tokenized sentences (lists of tokens/words)
# references should be: list of lists of reference lists (one reference per sentence)
# candidates should be: list of lists (one candidate per sentence)

# Simple token-based setup
references = [
    ["the", "quick", "brown", "fox"],
    ["a", "lazy", "dog"]
]

candidates = [
    ["the", "fast", "brown", "fox"],
    ["a", "sleepy", "dog"]
]

ARGS = {
    "references": references,
    "candidates": candidates
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk, 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
calculate_meteor_score Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/beta_code/machine_metrics.py
code served (permissive licence) · get_code("de1c559ab62ebfd7")
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.

from statistics import mean
from nltk.translate.meteor_score import meteor_score

# The function expects:
# - references: a list of reference sentences (each sentence is a list of words)
# - candidates: a list of candidate sentences (each sentence is a list of words)

# Looking at the function body:
# - references is iterated over: [[[j for j in i]] for i in references]
#   This means references should be a list of iterables (strings or lists of chars/words)
# - candidates is iterated over: [[j for j in i] for i in candidates]
#   This means candidates should be a list of iterables (strings or lists of chars/words)
# - meteor_score(ref, hyp) expects ref to be a list of reference translations
#   and hyp to be a hypothesis translation (list of words)

# NLTK's meteor_score signature: meteor_score(reference_corpus, hypothesis_sentence)
# where reference_corpus is a list of reference sentences (each is a list of words)
# and hypothesis_sentence is a list of words

ARGS = {
    "references": [
        "the quick brown fox",
        "a fast brown fox"
    ],
    "candidates": [
        "the quick brown fox",
        "a fast brown fox"
    ]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk — 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
correlation Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/beta_code/indicator_correlation.py
code served (permissive licence) · get_code("0d27332b2704f923")
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
from collections import defaultdict

# Mock the human_metrics module since it's imported by the function
import sys
from types import ModuleType

mock_human_metrics = ModuleType('human_metrics')
mock_human_metrics.mode_ref_dict = {
    2: 'model_2',
    3: 'model_3',
    4: 'model_4',
}
sys.modules['human_metrics'] = mock_human_metrics

# Construct model_scores
model_scores = {
    'prompt_1': {
        2: {
            'bleu_1': 0.45,
            'bleu_2': 0.38,
            'bleu_3': 0.32,
            'bleu_4': 0.28,
            'gleu': 0.40,
            'rouge_1': 0.50,
            'rouge_2': 0.42,
            'rouge_l': 0.48,
            'distinct_1': 0.85,
            'distinct_2': 0.72,
        },
        3: {
            'bleu_1': 0.52,
            'bleu_2': 0.44,
            'bleu_3': 0.38,
            'bleu_4': 0.33,
            'gleu': 0.48,
            'rouge_1': 0.58,
            'rouge_2': 0.50,
            'rouge_l': 0.55,
            'distinct_1': 0.88,
            'distinct_2': 0.76,
        },
    },
    'prompt_2': {
        2: {
            'bleu_1': 0.48,
            'bleu_2': 0.40,
            'bleu_3': 0.34,
            'bleu_4': 0.30,
            'gleu': 0.42,
            'rouge_1': 0.52,
            'rouge_2': 0.44,
            'rouge_l': 0.50,
            'distinct_1': 0.87,
            'distinct_2': 0.74,
        },
        3: {
            'bleu_1': 0.55,
            'bleu_2': 0.46,
            'bleu_3': 0.40,
            'bleu_4': 0.35,
            'gleu': 0.50,
            'rouge_1': 0.60,
            'rouge_2': 0.52,
            'rouge_l': 0.57,
            'distinct_1': 0.90,
            'distinct_2': 0.78,
        },
    },
}

# Construct human_scores
human_scores = {
    'prompt_1': {
        2: {
            'h_score': 3.5,
            'f_score': 3.2,
            'd_score': 3.8,
            'is_best': 0,
        },
        3: {
            'h_score': 4.1,
            'f_score': 3.9,
            'd_score': 4.2,
            'is_best': 1,
        },
    },
    'prompt_2': {
        2: {
            'h_score': 3.6,
            'f_score': 3.3,
            'd_score': 3.9,
            'is_best': 0,
        },
        3: {
            'h_score': 4.2,
            'f_score': 4.0,
            'd_score': 4.3,
            'is_best': 1,
        },
    },
}

ARGS = {
    'model_scores': model_scores,
    'human_scores': human_scores,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
human_metrics, numpy, scipy, 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
create_logger Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/cpm/generate_eval_data.py
code served (permissive licence) · get_code("a1b8c9fce8bf2bc6")
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 Logger. 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
import tempfile
import os

class Args:
    def __init__(self):
        # Create a temporary directory for the log file
        self.temp_dir = tempfile.mkdtemp()
        self.log_path = os.path.join(self.temp_dir, "test.log")

args_instance = Args()

ARGS = {
    "args": args_instance
}
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 Logger — a Logger output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Logger 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
func Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/beta_code/muti_heatmap_demo.py
code served (permissive licence) · get_code("97f913018bc28b19")
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

x = ['Cycle {}'.format(i) for i in range(1, 7)]

def func(x, pos):
    return '{:.2f}'.format(x).replace('0.', '.').replace('1.00', '')

ARGS = {
    'x': 0.42,
    'pos': 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
generate_text_by_input Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/cpm/generate_eval_data.py
code served (permissive licence) · get_code("668443ea1a908473")
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 torch.nn as nn
import torch.nn.functional as F

torch.manual_seed(0)

class MockTokenizer:
    def __init__(self):
        self.cls_token_id = 101
        self.bos_token_id = 102
        self.sep_token_id = 103
        self.pad_token_id = 0
        self.vocab_size = 1000
    
    def encode(self, text, add_special_tokens=False):
        return [10, 11, 12, 13, 14]
    
    def convert_tokens_to_ids(self, token):
        token_map = {
            '<unk>': 100,
            '<eop>': 104,
            '<eod>': 105,
            '▁。': 106,
            '▁?': 107,
            '▁!': 108,
            '▁(': 109,
            '▁0': 110,
            '0': 111,
            '▁1': 112,
            '▁2': 113,
            '▁3': 114,
            '▁4': 115,
            '▁5': 116,
            '▁6': 117,
            '▁7': 118,
            '▁8': 119,
            '▁9': 120,
            '▁.': 121,
            '.': 122,
        }
        return token_map.get(token, 200)
    
    def decode(self, token_ids):
        return "generated text sample"

class MockModel(nn.Module):
    def __init__(self, vocab_size=1000, hidden_dim=64):
        super().__init__()
        self.hidden_dim = hidden_dim
        self.vocab_size = vocab_size
        self.embedding = nn.Embedding(vocab_size, hidden_dim)
        self.transformer = nn.TransformerEncoderLayer(
            d_model=hidden_dim,
            nhead=2,
            dim_feedforward=128,
            batch_first=True,
            dropout=0.1
        )
        self.lm_head = nn.Linear(hidden_dim, vocab_size)
        self.device = 'cpu'
    
    def forward(self, input_ids):
        embeddings = self.embedding(input_ids)
        transformer_out = self.transformer(embeddings)
        logits = self.lm_head(transformer_out)
        return type('Output', (), {'logits': logits})()

class MockArgs:
    def __init__(self):
        self.utterance_max_len = 20
        self.seq_max_len = 50
        self.repetition_penalty = 1.2
        self.diverse_penalty = 1.1
        self.temperature = 0.9
        self.topk = 10
        self.topp = 0.95

model = MockModel(vocab_size=1000, hidden_dim=64)
tokenizer = MockTokenizer()
args = MockArgs()

ARGS = {
    'args': args,
    'text': 'Hello',
    'history': ['How are you?'],
    'model': model,
    'tokenizer': tokenizer,
}
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
heatmap Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/beta_code/muti_heatmap_demo.py
code served (permissive licence) · get_code("a9865f96feffb830")
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 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 numpy as np
import matplotlib.pyplot as plt

# Create a small 2D numpy array (4 rows x 3 columns for distinct dimensions)
data = np.random.randn(4, 3)

# Create row labels (4 labels for 4 rows)
row_labels = ['Row1', 'Row2', 'Row3', 'Row4']

# Create column labels (3 labels for 3 columns)
col_labels = ['Col1', 'Col2', 'Col3']

# Create a matplotlib axes object
fig, ax = plt.subplots()

# Create colorbar keyword arguments
cbar_kw = {'shrink': 0.8}

# Create colorbar label
cbarlabel = 'Value'

# Additional kwargs for imshow (like cmap, vmin, vmax)
imshow_kwargs = {'cmap': 'viridis'}

ARGS = {
    'data': data,
    'row_labels': row_labels,
    'col_labels': col_labels,
    'ax': ax,
    'cbar_kw': cbar_kw,
    'cbarlabel': cbarlabel,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
matplotlib, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
single_model_print Ran
ran on a synthesized input
anonNo2/crosstalk-generation/src/beta_code/matlibplot_heatmap.py
code served (permissive licence) · get_code("db65515177159aa1")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 3, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np
from scipy import stats

line_head = ['bleu_1', 'bleu_2', 'bleu_3', 'bleu_4', 'gleu', 'rouge_1', 'rouge_2', 'rouge_l', 'distinct_1', 'distinct_2']

col_head = ['h_score', 'f_score', 'is_best', 'd_score']

model_desc = [
    ('h_score', 'bleu_1', 0.85, 0.87),
    ('f_score', 'bleu_2', 0.72, 0.74),
    ('is_best', 'rouge_1', 0.91, 0.89),
    ('d_score', 'distinct_1', 0.65, 0.67),
    ('h_score', 'gleu', 0.78, 0.80),
    ('f_score', 'rouge_2', 0.68, 0.70),
]

ARGS = {
    'model_desc': model_desc
}
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 tuple, ndim 3, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
top_k_top_p_filtering Ran
ran; our contract draft was wrong, not the code
anonNo2/crosstalk-generation/src/cpm/generate_eval_data.py
code served (permissive licence) · get_code("6f5ddea9dd9cffbb")
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 1, dtype family float, shape (10,) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import torch.nn.functional as F

torch.manual_seed(0)

# Create a small logits tensor with vocabulary size
# The function expects logits to be 1D (vocab_size,)
logits = torch.randn(10)

# Set up parameters for top_k_top_p_filtering
ARGS = {
    "logits": logits,
    "top_k": 5,
    "top_p": 0.9,
    "filter_value": -float('Inf')
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 1, dtype family float, shape (10,) — shape (10,) 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
scatter Not yet run
unverified (harvested, no recorded run)
anonNo2/crosstalk-generation/src/gpt/data_parallel.py
code served (permissive licence) · get_code("69598ced7806d8c3")
How it ran

no invocation recorded.

scatter_kwargs Not yet run
unverified (harvested, no recorded run)
anonNo2/crosstalk-generation/src/gpt/data_parallel.py
code served (permissive licence) · get_code("aa1de448b0528802")
How it ran

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, 4 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 3 later) and 9 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
scatter_kwargs 11 other papers
3 dated earlier, 8 later
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
2106.06295 (2021-06)
Going Beyond Linear Transformers with Recurrent Fast Weight Programmers
2204.03883 (2022-04)
Vision Transformers for Single Image Dehazing
and 8 more not named here
scatter 8 other papers
3 dated earlier, 5 later
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
2106.06295 (2021-06)
Going Beyond Linear Transformers with Recurrent Fast Weight Programmers
2204.03883 (2022-04)
Vision Transformers for Single Image Dehazing
and 5 more not named here
top_k_top_p_filtering 4 other papers
4 dated earlier
1908.05739 (2019-08)
Abductive Commonsense Reasoning
1910.10683 (2019-10)
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
2005.00558 (2020-05)
POINTER: Constrained Progressive Text Generation via Insertion-based Generative Pre-training
and 1 more not named here
annotate_heatmap 1 other paper
1 later
2412.09115 (2024-12)
Vision CNNs trained to estimate spatial latents learned similar ventral-stream-aligned representations

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

Language is the principal tool for human communication, in which humor is one of the most attractive parts. Producing natural language like humans using computers, a.k.a, Natural Language Generation (NLG), has been widely used for dialogue systems, chatbots, machine translation, as well as computer-aid creation e.g., idea generations, scriptwriting. However, the humor aspect of natural language is relatively under-investigated, especially in the age of pre-trained language models. In this work, we aim to preliminarily test whether NLG can generate humor as humans do. We build a new dataset consisting of numerous digitized Chinese Comical Crosstalk scripts (called C 3 in short), which is for a popular Chinese performing art called 'Xiangsheng' or '相声' since 1800s 1 . We benchmark various generation approaches including training-from-scratch Seq2seq, fine-tuned middle-scale PLMs, and large-scale PLMs (with and without fine-tuning). Moreover, we also conduct a human assessment, showing that 1) large-scale pretraining largely improves crosstalk generation quality; and 2) even the scripts generated from the best PLM is far from what we expect, with only 65% quality of human-created crosstalk. We conclude, humor generation could be largely improved using large-scaled PLMs, but it is still in its infancy. The data and benchmarking code is publicly available in https://github.com/anonNo2/crosstalk-generation.

For agents

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

get_harvested_code_for_paper("2207.00735")
get_code_for_paper("2207.00735")
have("2207.00735")

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

curl https://syntology.ai/api/ran/2207.00735.json

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

[![Syntology run record](https://syntology.ai/api/ran/2207.00735.svg)](https://syntology.ai/paper/2207.00735)

Connect an agent — have() is free.