SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2404.19563 · EMNLP · 2024

RepEval: Effective Text Evaluation with LLM Representation

Yi Xu, Chenghu Zhou, Xinbing Wang, Xiaoying Gan, Lei Zhou, Jiaxin Ding, Tianhang Zhang, Luoyi Fu, Shuqian Sheng, Zanwei Shen

arXiv · PDF · Open in the Atlas

Code that ran

18 of 27 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; 4 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 14 failures of our instrument; 9 no recorded run); 1 of them was also run on the host, not in the jail, by the peer-agreement lane that set its level, 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/2404.19563.json · badge: /api/ran/2404.19563.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
shikib/usr canonical (classifier label)
classifier basis: census
10 ran · 2 no recorded run
maszhongming/UniEval canonical (classifier label)
classifier basis: census
8 ran · 3 no recorded run
jinlanfu/GPTScore community
Papers with Code archive
0 ran · 3 no recorded run
neulab/BARTScore community
Papers with Code archive
0 ran · 1 no recorded run
FunctionStatusWhere it lives
swish Ran
ran; our contract draft was wrong, not the code
shikib/usr/transformers/modeling_openai.py
code served (permissive licence) · get_code("0f786c407fb1ee4c")
How it ran

Ran in Syntology's jail on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. The comparison behind this sample's level is a different run: the peer-agreement lane ran it again on the host, not in the jail (report dated 2026-09-04T22:33:52Z). 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

torch.manual_seed(0)

# The function swish takes a single parameter 'x' which is used with:
# - torch.sigmoid(x) - requires a tensor
# - x * ... - requires a tensor that supports multiplication
# So x should be a tensor

x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
add_question Ran
ran; our contract draft was wrong, not the code
maszhongming/UniEval/utils.py
code served (permissive licence) · get_code("1a5cb15d63f16098")
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

torch.manual_seed(0)

ARGS = {
    'dimension': 'fluency',
    'output': [
        'This is a fluent paragraph.',
        'Another fluent text example.'
    ],
    'src': [
        'The source document for summarization task.',
        'Another source document here.'
    ],
    'ref': [
        'The human-annotated reference text.',
        'Another reference example.'
    ],
    'context': [
        'Additional factual information for context.',
        'More context information here.'
    ],
    'task': 'summarization'
}
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
angle_defn Ran
ran; our contract draft was wrong, not the code
shikib/usr/transformers/modeling_ctrl.py
code served (permissive licence) · get_code("671a619bed5669a3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# pos: should be a tensor representing positions
# The function multiplies pos by angle_rates, so pos should be numeric
pos = torch.arange(0, 2, dtype=torch.float32).unsqueeze(1)  # shape: (2, 1)

# i: index used in the calculation (2 * (i // 2) / d_model_size)
# Should be an integer or tensor of integers
i = torch.tensor([0, 1, 2, 3], dtype=torch.long)

# d_model_size: the model dimension size, used as divisor
# Should be a positive integer
d_model_size = 4

ARGS = {
    "pos": pos,
    "i": i,
    "d_model_size": d_model_size
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
calculate_correlation Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
maszhongming/UniEval/reproduce/correlation.py
code served (permissive licence) · get_code("31b107371370f1bf")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np
from scipy.stats import spearmanr, pearsonr, kendalltau

torch.manual_seed(0)

# Create small prediction scores
pred_score = [0.1, 0.2, 0.3, 0.4, 0.5]

# Create small human scores
human_score = [0.15, 0.25, 0.35, 0.45, 0.55]

# Create dimension identifier
dim = "overall_quality"

# Create empty result dictionary to store correlation metrics
result = {}

ARGS = {
    "pred_score": pred_score,
    "human_score": human_score,
    "dim": dim,
    "result": result
}
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 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
convert_to_json Ran
ran; our contract draft was wrong, not the code
maszhongming/UniEval/utils.py
code served (permissive licence) · get_code("40a6d254f3b50b99")
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

torch.manual_seed(0)

# Construct arguments for convert_to_json function
output_list = [
    "This is the model output for sample 1.",
    "This is the model output for sample 2."
]

src_list = [
    "This is the source document for sample 1.",
    "This is the source document for sample 2."
]

ref_list = [
    "This is the reference/groundtruth for sample 1.",
    "This is the reference/groundtruth for sample 2."
]

context_list = [
    "This is additional context for sample 1.",
    "This is additional context for sample 2."
]

scores = [
    {'fluency': 2.5, 'coherence': 3.0},
    {'fluency': 3.5, 'coherence': 2.5}
]

doc_id = [0, 1]

system_id = [0, 1]

ARGS = {
    'output_list': output_list,
    'src_list': src_list,
    'ref_list': ref_list,
    'context_list': context_list,
    'scores': scores,
    'doc_id': doc_id,
    'system_id': system_id
}
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 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
gelu Ran
ran; our contract draft was wrong, not the code
shikib/usr/transformers/modeling_gpt2.py
code served (permissive licence) · get_code("8d23fbe2b99b840b")
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 Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import math

torch.manual_seed(0)

# Create a small tensor for testing
x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
gelu Ran
ran, honoured the contract we drafted
shikib/usr/transformers/modeling_bert.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from lonePatient/BERT-SDA, is pointer only (licence: NONE) · get_code("211753ba29188d4e")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from lonePatient/BERT-SDA, 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.

gelu Ran
ran; our contract draft was wrong, not the code
shikib/usr/transformers/modeling_distilbert.py
code served (permissive licence) · get_code("69c6d14de8190cfb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import math
import torch

torch.manual_seed(0)

# The function gelu takes a single parameter x which is used with:
# - multiplication operator (*)
# - torch.erf() function
# - division operator (/)
# This suggests x should be a torch.Tensor

x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 4) — shape (2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
gelu_new Ran
ran, honoured the contract we drafted
shikib/usr/transformers/modeling_bert.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from lonePatient/BERT-SDA, is pointer only (licence: NONE) · get_code("77601724cad03f95")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from lonePatient/BERT-SDA, 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_unique_value Ran
ran, honoured the contract we drafted
maszhongming/UniEval/reproduce/correlation.py
code served (permissive licence) · get_code("729c770d517b1bc6")
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

torch.manual_seed(0)

# Construct data as a list of dictionaries
data = [
    {"name": "Alice", "age": 30, "city": "NYC"},
    {"name": "Bob", "age": 25, "city": "LA"},
    {"name": "Charlie", "age": 30, "city": "NYC"},
    {"name": "Diana", "age": 25, "city": "Chicago"},
]

key = "age"

ARGS = {
    "data": data,
    "key": key,
}
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
load_data Ran
ran; our contract draft was wrong, not the code
maszhongming/UniEval/pseudo_data_summ.py
code served (permissive licence) · get_code("cdc21dc661030083")
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 json
import tempfile
import os

# Create a temporary directory and a mock JSONL file
temp_dir = tempfile.mkdtemp()
data_path = os.path.join(temp_dir, 'cnndm_train.jsonl')

# Write some sample data to the JSONL file
sample_data = [
    {"text": "This is the first document.", "summary": "First doc"},
    {"text": "This is the second document.", "summary": "Second doc"},
    {"text": "This is the third document.", "summary": "Third doc"}
]

with open(data_path, 'w') as f:
    for item in sample_data:
        f.write(json.dumps(item) + '\n')

ARGS = {
    'data_path': data_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 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
load_json Ran
ran; our contract draft was wrong, not the code
maszhongming/UniEval/reproduce/data_utils.py
code served (permissive licence) · get_code("79ec7dfe27ceef16")
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 json
import tempfile
import os

# Create a temporary JSON file for testing
temp_dir = tempfile.mkdtemp()
data_path = os.path.join(temp_dir, "test_data.json")

# Write sample JSON data to the file
sample_data = {
    "key1": "value1",
    "key2": [1, 2, 3],
    "key3": {"nested": "object"}
}

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

ARGS = {
    "data_path": data_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 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
load_output Ran
ran; our contract draft was wrong, not the code
maszhongming/UniEval/billboard/evaluate.py
code served (permissive licence) · get_code("154c40d3e5f52366")
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 json
import tempfile
import os

# Create a temporary directory and file for the test
temp_dir = tempfile.mkdtemp()
jsonl_file = os.path.join(temp_dir, 'generator-output.jsonl')

# Write sample JSONL data
with open(jsonl_file, 'w') as f:
    f.write(json.dumps({'hyp': 'hypothesis 1'}) + '\n')
    f.write(json.dumps({'hyp': 'hypothesis 2'}) + '\n')
    f.write(json.dumps({'hyp': 'hypothesis 3'}) + '\n')

# Change to the temp directory so the function can find 'generator-output.jsonl'
original_dir = os.getcwd()
os.chdir(temp_dir)

ARGS = {
    'output_path': jsonl_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, 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
load_src Ran
ran; our contract draft was wrong, not the code
maszhongming/UniEval/billboard/evaluate.py
code served (permissive licence) · get_code("7f6c2b0cc66bbc7c")
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 json
import tempfile
import os

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

# Write sample JSONL data with 'src' field
sample_data = [
    {'src': 'hello world'},
    {'src': 'test source'},
    {'src': 'another example'},
]

for item in sample_data:
    temp_file.write(json.dumps(item) + '\n')

temp_file.close()

ARGS = {
    'src_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 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
load_tf_weights_in_bert Ran
ran; our contract draft was wrong, not the code
shikib/usr/transformers/modeling_bert.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from tanmay1618/text_to_sql_bert, is pointer only (licence: NONE) · get_code("0c145728a35789fa")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from tanmay1618/text_to_sql_bert, 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.

positional_encoding Ran
ran, honoured the contract we drafted
shikib/usr/transformers/modeling_ctrl.py
code served (permissive licence) · get_code("59ac77747a12b1cb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (8, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

ARGS = {
    "position": 8,
    "d_model_size": 4,
    "dtype": torch.float32
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (8, 4) — shape (8, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
scaled_dot_product_attention Ran
ran; our contract draft was wrong, not the code
shikib/usr/transformers/modeling_ctrl.py
code served (permissive licence) · get_code("57ebc1447d921c46")
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)

# Construct tiny tensors for q, k, v
# q shape: (batch=2, num_heads=2, seq_len_q=3, d_k=4)
# k shape: (batch=2, num_heads=2, seq_len_k=3, d_k=4)
# v shape: (batch=2, num_heads=2, seq_len_v=3, d_v=5)

q = torch.randn(2, 2, 3, 4)
k = torch.randn(2, 2, 3, 4)
v = torch.randn(2, 2, 3, 5)

# mask: optional binary mask, same shape as attention logits
# attention logits shape will be (2, 2, 3, 3) after matmul(q, k.T)
mask = torch.zeros(2, 2, 3, 3)

# attention_mask: optional additive mask, same shape as attention logits
attention_mask = torch.zeros(2, 2, 3, 3)

# head_mask: optional mask for heads, shape (num_heads,) or broadcastable
head_mask = torch.ones(2, 2, 1, 1)

ARGS = {
    "q": q,
    "k": k,
    "v": v,
    "mask": mask,
    "attention_mask": attention_mask,
    "head_mask": head_mask,
}
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
tile Ran
ran; our contract draft was wrong, not the code
shikib/usr/transformers/modeling_beam_search.py
code served (permissive licence) · get_code("a878dc3112f9ce9e")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (6, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small tensor with distinct dimensions
# Using shape (2, 4) - batch size 2, feature dim 4
x = torch.tensor([[1, 2, 3, 4], [5, 6, 7, 8]], dtype=torch.float32)

# count: number of times to tile
count = 3

# dim: dimension along which to tile
dim = 0

ARGS = {
    "x": x,
    "count": count,
    "dim": dim,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (6, 4) — shape (6, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
BARTScorer Not yet run
unverified (harvested, no recorded run)
neulab/BARTScore/bart_score.py
code served (permissive licence) · get_code("3186cad3f4fa28a4")
How it ran

no invocation recorded.

GPT3Model Not yet run
unverified (harvested, no recorded run)
jinlanfu/GPTScore/score_d2t.py
pointer only (licence: NONE) · get_code("388a367eded0b858")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinlanfu/GPTScore 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.

Scorer Not yet run
unverified (harvested, no recorded run)
jinlanfu/GPTScore/score_d2t.py
pointer only (licence: NONE) · get_code("70093442b3ee191f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinlanfu/GPTScore 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.

disfluency_transformation Not yet run
unverified (harvested, no recorded run)
maszhongming/UniEval/pseudo_data_summ.py
code served (permissive licence) · get_code("666cfad4555e9faa")
How it ran

no invocation recorded.

gpt3score Not yet run
unverified (harvested, no recorded run)
jinlanfu/GPTScore/score_d2t.py
pointer only (licence: NONE) · get_code("6b02658358cb6a62")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinlanfu/GPTScore 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.

incoherence_transformation Not yet run
unverified (harvested, no recorded run)
maszhongming/UniEval/pseudo_data_summ.py
code served (permissive licence) · get_code("bb6e70aa3a9fcf4e")
How it ran

no invocation recorded.

load_ref Not yet run
unverified (harvested, no recorded run)
maszhongming/UniEval/billboard/evaluate.py
code served (permissive licence) · get_code("dbd9587f7a986fc3")
How it ran

no invocation recorded.

load_tf_weights_in_gpt2 Not yet run
unverified (harvested, no recorded run)
shikib/usr/transformers/modeling_gpt2.py
code served (permissive licence) · get_code("654e91efc679fe84")
How it ran

no invocation recorded.

load_tf_weights_in_openai_gpt Not yet run
unverified (harvested, no recorded run)
shikib/usr/transformers/modeling_openai.py
code served (permissive licence) · get_code("42ee168c3ae12e60")
How it ran

no invocation recorded.

Also present in other papers' code

Of 27 samples lifted from this paper's linked repositories, 27 carry a function identity and 0 do not. Of the 27 distinct functions with an identity, 13 are also present in other papers whose code we hold (13 with at least one paper dated earlier, 3 the same month, 5 later) and 14 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
swish 206 other papers
178 dated earlier, 1 the same month, 23 later, 4 same year, order not determinable · 44 only as an assembly brick
1408.5882 (2014-08)
Convolutional Neural Networks for Sentence Classification
1805.12114 (2018-05)
Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models
1809.01357 (2018-09)
Zero Shot Learning for Code Education: Rubric Sampling with Deep Learning Inference
and 203 more not named here
gelu 198 other papers
167 dated earlier, 1 the same month, 27 later, 3 same year, order not determinable · 32 only as an assembly brick
1408.5882 (2014-08)
Convolutional Neural Networks for Sentence Classification
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.10125 (2019-01)
Glyce: Glyph-vectors for Chinese Character Representations
and 195 more not named here
gelu 51 other papers
48 dated earlier, 2 later, 1 same year, order not determinable · 8 only as an assembly brick
1706.03762 (2017-06)
Attention Is All You Need
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
and 48 more not named here
gelu_new 23 other papers
22 dated earlier, 1 later
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1909.11942 (2019-09)
ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
1911.03631 (2019-11)
Hierarchical Graph Network for Multi-hop Question Answering
and 20 more not named here
load_tf_weights_in_bert 19 other papers
18 dated earlier, 1 same year, order not determinable
1810.04805 (2018-10)
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
1902.01069 (2019-02)
A Comprehensive Exploration on WikiSQL with Table-Aware Word Contextualization
1909.11942 (2019-09)
ALBERT: A Lite BERT for Self-supervised Learning of Language Representations
and 16 more not named here
gelu 17 other papers
14 dated earlier, 1 the same month, 1 later, 1 same year, order not determinable
1905.12790 (2019-05)
A Generalized Framework of Sequence Generation with Application to Undirected Sequence Models
1906.08237 (2019-06)
XLNet: Generalized Autoregressive Pretraining for Language Understanding
1909.05311 (2019-09)
Graph-Based Reasoning over Heterogeneous External Knowledge for Commonsense Question Answering
and 14 more not named here
load_tf_weights_in_openai_gpt 8 other papers
7 dated earlier, 1 same year, order not determinable
1906.08237 (2019-06)
XLNet: Generalized Autoregressive Pretraining for Language Understanding
1909.05311 (2019-09)
Graph-Based Reasoning over Heterogeneous External Knowledge for Commonsense Question Answering
2002.01808 (2020-02)
K-Adapter: Infusing Knowledge into Pre-Trained Models with Adapters
and 5 more not named here
tile 8 other papers
8 dated earlier
1905.13164 (2019-05)
Hierarchical Transformers for Multi-Document Summarization
1906.10508 (2019-06)
Non-Parallel Sequence-to-Sequence Voice Conversion with Disentangled Linguistic and Speaker Representations
2001.09099 (2020-01)
TVR: A Large-Scale Dataset for Video-Subtitle Moment Retrieval
and 5 more not named here
load_tf_weights_in_gpt2 6 other papers
5 dated earlier, 1 same year, order not determinable
2002.01808 (2020-02)
K-Adapter: Infusing Knowledge into Pre-Trained Models with Adapters
2004.03829 (2020-04)
Exploring Versatile Generative Language Model Via Parameter-Efficient Transfer Learning
2010.05607 (2020-10)
The elephant in the interpretability room: Why use attention as explanation when we have saliency methods?
and 3 more not named here
angle_defn 2 other papers
2 dated earlier
2021.acl-long.390 (2021 (year only))
Few-Shot Text Ranking with Meta Adapted Synthetic Weak Supervision
2212.03506 (2022-12)
WIDER & CLOSER: Mixture of Short-channel Distillers for Zero-shot Cross-lingual Named Entity Recognition
positional_encoding 2 other papers
2 dated earlier
2021.acl-long.390 (2021 (year only))
Few-Shot Text Ranking with Meta Adapted Synthetic Weak Supervision
2212.03506 (2022-12)
WIDER & CLOSER: Mixture of Short-channel Distillers for Zero-shot Cross-lingual Named Entity Recognition
BARTScorer
brick of this paper's assembly: entry
1 other paper
1 dated earlier
2211.01635 (2022-11)
Revisiting Grammatical Error Correction Evaluation and Beyond
scaled_dot_product_attention 1 other paper
1 dated earlier
2212.03506 (2022-12)
WIDER & CLOSER: Mixture of Short-channel Distillers for Zero-shot Cross-lingual Named Entity Recognition

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

The era of Large Language Models (LLMs) raises new demands for automatic evaluation metrics, which should be adaptable to various application scenarios while maintaining low cost and effectiveness. Traditional metrics for automatic text evaluation are often tailored to specific scenarios, while LLM-based evaluation metrics are costly, requiring fine-tuning or rely heavily on the generation capabilities of LLMs. Besides, previous LLM-based metrics ignore the fact that, within the space of LLM representations, there exist direction vectors that indicate the estimation of text quality. To this end, we introduce RepEval, a metric that leverages the projection of LLM representations for evaluation. Through simple prompt modifications, RepEval can easily transition to various tasks, requiring only minimal sample pairs for direction vector construction. Results on fourteen datasets across two evaluation tasks demonstrate the high effectiveness of our method, which exhibits a higher correlation with human judgments than previous methods, even in complex evaluation scenarios involving pair-wise selection under nuanced aspects. Our work underscores the richness of information regarding text quality embedded within LLM representations, offering insights for the development of new metrics. 1 *

For agents

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

get_harvested_code_for_paper("2404.19563")
get_code_for_paper("2404.19563")
have("2404.19563")

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

curl https://syntology.ai/api/ran/2404.19563.json

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

[![Syntology run record](https://syntology.ai/api/ran/2404.19563.svg)](https://syntology.ai/paper/2404.19563)

Connect an agent — have() is free.