SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.01574 · NeurIPS · 2024

MMLU-Pro: A More Robust and Challenging Multi-Task Language Understanding Benchmark

Wenhu Chen, Kai Wang, Yubo Wang, Ge Zhang, Abhranil Chandra, Xueguang Ma, Yuansheng Ni, Shiguang Guo, Weiming Ren, Aaran Arulraj, Xuan He, Ziyan Jiang, and 5 more

arXiv · PDF · Open in the Atlas

Code that ran

12 of 12 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 3 have no contract verdict; 9 failures of our instrument; 0 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/2406.01574.json · badge: /api/ran/2406.01574.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
tiger-ai-lab/mmlu-pro canonical (classifier label)
classifier basis: classifier_2025
5 ran · 0 no recorded run
wenlai-lavine/jola community (classifier label)
classifier basis not recorded on the link
7 ran · 0 no recorded run
FunctionStatusWhere it lives
repeat_kv Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
wenlai-lavine/jola/jola/modeling_llama.py
code served (permissive licence) · get_code("30d7eec482ebf6b1")
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 4, dtype family float, shape (2, 6, 5, 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 2026-09-04T23:09:20Z). 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 tensor with shape (batch, num_key_value_heads, seqlen, head_dim)
# Using distinct sizes: batch=2, num_key_value_heads=3, seqlen=5, head_dim=4
hidden_states = torch.randn(2, 3, 5, 4)

# n_rep is the repetition factor
n_rep = 2

ARGS = {
    "hidden_states": hidden_states,
    "n_rep": n_rep,
}
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 4, dtype family float, shape (2, 6, 5, 4) — shape (2, 6, 5, 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
apply_rotary_pos_emb Ran
ran; our contract draft was wrong, not the code
wenlai-lavine/jola/jola/modeling_qwen2.py
code served (permissive licence) · get_code("bac65c3dafaec040")
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 tuple, ndim 5, 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)

# Create small tensors for query and key
# Shape: [batch_size, heads, seq_len, head_dim]
batch_size = 2
heads = 3
seq_len = 4
head_dim = 8

q = torch.randn(batch_size, heads, seq_len, head_dim)
k = torch.randn(batch_size, heads, seq_len, head_dim)

# Create cos and sin tensors
# Shape: [batch_size, seq_len, head_dim]
cos = torch.randn(batch_size, seq_len, head_dim)
sin = torch.randn(batch_size, seq_len, head_dim)

# position_ids is optional and unused according to docstring
position_ids = None

# unsqueeze_dim defaults to 1
unsqueeze_dim = 1

ARGS = {
    "q": q,
    "k": k,
    "cos": cos,
    "sin": sin,
    "position_ids": position_ids,
    "unsqueeze_dim": unsqueeze_dim,
}
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 tuple, ndim 5, 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
apply_rotary_pos_emb Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
wenlai-lavine/jola/jola/modeling_llama.py
code served (permissive licence) · get_code("f725bc2d76076485")
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 tuple, ndim 5, 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)

# Create small tensors for q and k
# Shape: (batch_size=2, seq_len=3, hidden_dim=4)
q = torch.randn(2, 3, 4)
k = torch.randn(2, 3, 4)

# Create cos and sin tensors
# The function squeezes dimensions 1 and 0, then indexes by position_ids
# So we need shape that allows this: (1, 1, seq_len, hidden_dim)
cos = torch.randn(1, 1, 5, 4)
sin = torch.randn(1, 1, 5, 4)

# Create position_ids
# Shape: (batch_size=2, seq_len=3)
position_ids = torch.tensor([[0, 1, 2], [1, 2, 3]])

ARGS = {
    "q": q,
    "k": k,
    "cos": cos,
    "sin": sin,
    "position_ids": position_ids,
}
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 tuple, ndim 5, 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
extract_again Ran
ran; our contract draft was wrong, not the code
tiger-ai-lab/mmlu-pro/compute_accuracy.py
code served (permissive licence) · get_code("628ba40cfdad9e0a")
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 re

def extract_final(text):
    pattern = '\\b[A-J]\\b(?!.*\\b[A-J]\\b)'
    match = re.search(pattern, text, re.DOTALL)
    if match:
        return match.group(0)
    else:
        return None

def extract_again(text):
    match = re.search('.*[aA]nswer:\\s*([A-J])', text)
    if match:
        return match.group(1)
    else:
        return extract_final(text)

# Test case 1: text with "Answer:" followed by a letter
text1 = "Some question here\nAnswer: C"

# Test case 2: text with "answer:" followed by a letter
text2 = "Some question\nanswer: F\nMore text"

# Test case 3: text with multiple letters but no Answer field - should extract final
text3 = "Question with letters A B C D E"

# Test case 4: text with Answer field containing lowercase answer
text4 = "Question\nAnswer: G"

ARGS = {
    "text": text1
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_answer Ran
ran; our contract draft was wrong, not the code
tiger-ai-lab/mmlu-pro/compute_accuracy.py
code served (permissive licence) · get_code("1f59ab74392c3fcf")
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 re

def extract_again(text):
    match = re.search('.*[aA]nswer:\\s*([A-J])', text)
    if match:
        return match.group(1)
    else:
        return extract_final(text)

def extract_final(text):
    pattern = '\\b[A-J]\\b(?!.*\\b[A-J]\\b)'
    match = re.search(pattern, text, re.DOTALL)
    if match:
        return match.group(0)
    else:
        return None

def extract_answer(text, level):
    if level == 'l1':
        pattern = 'answer is \\(?([A-J])\\)?'
        match = re.search(pattern, text)
        if match:
            return match.group(1)
        else:
            return None
    elif level == 'l2':
        pattern = 'answer is \\(?([A-J])\\)?'
        match = re.search(pattern, text)
        if match:
            return match.group(1)
        else:
            return extract_again(text)

ARGS = {
    'text': 'The answer is (C) according to the calculation.',
    'level': 'l1'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_common_reason_answer Ran
ran on a synthesized input
wenlai-lavine/jola/jola/evaluate.py
code served (permissive licence) · get_code("eaac056d48deb699")
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 re

def extract_common_reason_answer(sentence: str, task) -> float:
    if task == 'BoolQ':
        sentence_ = sentence.strip()
        pred_answers = re.findall('true|false', sentence_)
        if not pred_answers:
            return ''
        return pred_answers[0]
    elif task == 'PIQA':
        sentence_ = sentence.strip()
        pred_answers = re.findall('solution1|solution2', sentence_)
        if not pred_answers:
            return ''
        return pred_answers[0]
    elif task in ['SIQA', 'ARC-c', 'ARC-e', 'OBQA']:
        sentence_ = sentence.strip()
        pred_answers = re.findall('answer1|answer2|answer3|answer4|answer5', sentence_)
        if not pred_answers:
            return ''
        return pred_answers[0]
    elif task == 'HellaSwag':
        sentence_ = sentence.strip()
        pred_answers = re.findall('ending1|ending2|ending3|ending4', sentence_)
        if not pred_answers:
            return ''
        return pred_answers[0]
    elif task == 'WinoGrande':
        sentence_ = sentence.strip()
        pred_answers = re.findall('option1|option2', sentence_)
        if not pred_answers:
            return ''
        return pred_answers[0]

ARGS = {
    'sentence': 'The answer is true because it makes sense.',
    'task': 'BoolQ'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_final Ran
ran; our contract draft was wrong, not the code
tiger-ai-lab/mmlu-pro/compute_accuracy.py
code served (permissive licence) · get_code("bfcfbf7614a49849")
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 re

def extract_final(text):
    pattern = '\\b[A-J]\\b(?!.*\\b[A-J]\\b)'
    match = re.search(pattern, text, re.DOTALL)
    if match:
        return match.group(0)
    else:
        return None

# The function takes a single parameter 'text' which is a string
# It searches for the last occurrence of a word boundary followed by a single letter A-J followed by a word boundary
# Let's create a simple test case with multiple letters and ensure the last one is captured

text_input = "This is a test A with B multiple C letters D and E more F text G H I J"

ARGS = {
    'text': text_input
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_gem_answer Ran
ran on a synthesized input
wenlai-lavine/jola/jola/evaluate.py
code served (permissive licence) · get_code("a57a6eff6220ee2e")
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

torch.manual_seed(0)

ARGS = {
    "sentence": "This is a test sentence. With some punctuation. And more text."
}
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
extract_mmlu_pro_answer Ran
ran on a synthesized input
wenlai-lavine/jola/jola/evaluate.py
code served (permissive licence) · get_code("af54879256b59d48")
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 re

def extract_mmlu_pro_answer(sentence: str) -> float:
    sentence_ = sentence.strip()
    pred_answers = re.findall('A|B|C|D|E|F|G|H|I|J', sentence_)
    if not pred_answers:
        return ''
    return pred_answers[0]

# Setup code
ARGS = {
    'sentence': '  The answer is B because of the following reasons.  '
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
format_example Ran
ran; our contract draft was wrong, not the code
tiger-ai-lab/mmlu-pro/evaluate_from_api.py
code served (permissive licence) · get_code("1b3ff879d7f6d22b")
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

torch.manual_seed(0)

# Analyze the function:
# - question: used with .format(), expects a string
# - options: iterated over with enumerate(), expects an iterable of strings
# - cot_content: string with default '', checked with startswith() and ==, string slicing with [3:]

ARGS = {
    'question': 'What is the capital of France?',
    'options': ['Paris', 'London', 'Berlin', 'Madrid'],
    'cot_content': 'The capital of France is Paris because it is the largest city and the political center.'
}
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
preprocess Ran
ran; our contract draft was wrong, not the code
tiger-ai-lab/mmlu-pro/evaluate_from_api.py
code served (permissive licence) · get_code("c48a904c128bab0a")
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

# Construct test_df - a list of dictionaries with 'options' and 'category' keys
test_df = [
    {
        'options': ['option1', 'N/A', 'option2'],
        'category': 'category_A',
        'id': 1
    },
    {
        'options': ['N/A', 'option3'],
        'category': 'category_B',
        'id': 2
    },
    {
        'options': ['option4', 'option5', 'N/A'],
        'category': 'category_A',
        'id': 3
    }
]

ARGS = {
    'test_df': test_df
}
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 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
rotate_half Ran
ran; our contract draft was wrong, not the code
wenlai-lavine/jola/jola/modeling_llama.py
code served (permissive licence) · get_code("b99eea6376d1e212")
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

torch.manual_seed(0)

# The function rotate_half takes a single parameter 'x' which is a tensor
# It accesses x.shape[-1] to get the last dimension
# It uses indexing with [..., :] slicing notation
# It calls torch.cat() to concatenate tensors

# Create a small tensor with shape (2, 4) - batch size 2, hidden dim 4
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

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 8 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 4 the same month, 8 later) and 4 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
rotate_half 360 other papers
135 dated earlier, 21 the same month, 195 later, 6 same year, order not determinable, 3 undated · 30 only as an assembly brick
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
2203.08913 (2022-03)
Memorizing Transformers
2302.13971 (2023-02)
LLaMA: Open and Efficient Foundation Language Models
and 357 more not named here
repeat_kv 243 other papers
79 dated earlier, 16 the same month, 144 later, 2 same year, order not determinable, 2 undated · 22 only as an assembly brick
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
2203.08913 (2022-03)
Memorizing Transformers
2305.17888 (2023-05)
LLM-QAT: Data-Free Quantization Aware Training for Large Language Models
and 240 more not named here
apply_rotary_pos_emb 94 other papers
14 dated earlier, 6 the same month, 70 later, 1 same year, order not determinable, 3 undated · 6 only as an assembly brick
1901.02860 (2019-01)
Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context
2305.17888 (2023-05)
LLM-QAT: Data-Free Quantization Aware Training for Large Language Models
2309.11674 (2023-09)
A Paradigm Shift in Machine Translation: Boosting Translation Performance of Large Language Models
and 91 more not named here
apply_rotary_pos_emb 83 other papers
39 dated earlier, 6 the same month, 36 later, 2 same year, order not determinable · 6 only as an assembly brick
2306.15595 (2023-06)
Extending Context Window of Large Language Models via Positional Interpolation
2307.05695 (2023-07)
ReLoRA: High-Rank Training Through Low-Rank Updates
2307.06945 (2023-07)
In-context Autoencoder for Context Compression in a Large Language Model
and 80 more not named here
extract_again 2 other papers
2 later
2601.11866 (2026-01)
Advances in LLM Reasoning Enable Flexibility in Clinical Problem-Solving
2601.22139 (2026-01)
Reasoning While Asking: Transforming Reasoning Large Language Models from Passive Solvers to Proactive Inquirers
extract_final 2 other papers
2 later
2601.11866 (2026-01)
Advances in LLM Reasoning Enable Flexibility in Clinical Problem-Solving
2601.22139 (2026-01)
Reasoning While Asking: Transforming Reasoning Large Language Models from Passive Solvers to Proactive Inquirers
extract_answer 1 other paper
1 later
2601.11866 (2026-01)
Advances in LLM Reasoning Enable Flexibility in Clinical Problem-Solving
preprocess 1 other paper
1 later
2601.22139 (2026-01)
Reasoning While Asking: Transforming Reasoning Large Language Models from Passive Solvers to Proactive Inquirers

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

In the age of large-scale language models, benchmarks like the Massive Multitask Language Understanding (MMLU) have been pivotal in pushing the boundaries of what AI can achieve in language comprehension and reasoning across diverse domains. However, as models continue to improve, their performance on these benchmarks has begun to plateau, making it increasingly difficult to discern differences in model capabilities. This paper introduces MMLU-Pro, an enhanced dataset designed to extend the mostly knowledge-driven MMLU benchmark by integrating more challenging, reasoning-focused questions and expanding the choice set from four to ten options. Additionally, MMLU-Pro eliminates the trivial and noisy questions in MMLU. Our experimental results show that MMLU-Pro not only raises the challenge, causing a significant drop in accuracy by 16% to 33% compared to MMLU but also demonstrates greater stability under varying prompts. With 24 different prompt styles tested, the sensitivity of model scores to prompt variations decreased from 4-5% in MMLU to just 2% in MMLU-Pro. Additionally, we found that models utilizing Chain of Thought (CoT) reasoning achieved better performance on MMLU-Pro compared to direct answering, which is in stark contrast to the findings on the original MMLU, indicating that MMLU-Pro includes more complex reasoning questions. Our assessments confirm that MMLU-Pro is a more discriminative benchmark to better track progress in the field.

For agents

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

get_harvested_code_for_paper("2406.01574")
get_code_for_paper("2406.01574")
have("2406.01574")

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

curl https://syntology.ai/api/ran/2406.01574.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.01574.svg)](https://syntology.ai/paper/2406.01574)

Connect an agent — have() is free.