SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.20331 · 2024

Unsolvable Problem Detection: Robust Understanding Evaluation for Large Multimodal Models

arXiv · PDF · Open in the Atlas

Code that ran

7 of 9 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; 1 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 5 failures of our instrument; 2 no recorded run), measured 2026-11-21; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2403.20331.json · badge: /api/ran/2403.20331.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
atsumiyai/upd canonical (classifier label)
classifier basis: census
7 ran · 2 no recorded run
FunctionStatusWhere it lives
build_option_str_w_gt Ran
ran; our contract draft was wrong, not the code
atsumiyai/upd/automatic_eval/calculate_scores.py
code served (permissive licence) · get_code("e7c04c2111338866")
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 pandas as pd
import string
import math

# Build valid arguments for build_option_str_w_gt function

ARGS = {
    'option_list': ['Option A text', 'Option B text', None, None],
    'gt_text': 'This is the ground truth answer',
    'eval_type': 'aad',
    'question_type': 'inst',
    'upd_type': 'aad'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — 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
eval_result_dual Ran
ran; our contract draft was wrong, not the code
atsumiyai/upd/automatic_eval/utils/common_utils.py
code served (permissive licence) · get_code("189c31121c050c08")
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 pandas as pd
import numpy as np

# Create a small dataframe with the required columns
np.random.seed(0)

# Create sample data with 'split', 'hit', 'category', and 'l2-category' columns
n_samples = 10
data = {
    'split': ['test'] * n_samples,
    'hit': np.random.rand(n_samples),
    'category': np.random.choice(['cat_a', 'cat_b', 'cat_c'], n_samples),
    'l2-category': np.random.choice(['l2_x', 'l2_y', 'l2_z'], n_samples)
}
data_main = pd.DataFrame(data)

ARGS = {
    'data_main': data_main
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — 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
extract_options Ran
ran; our contract draft was wrong, not the code
atsumiyai/upd/automatic_eval/calculate_scores.py
code served (permissive licence) · get_code("6368891e7477c9bb")
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 pandas as pd
import math

def is_none(value):
    if value is None:
        return True
    if pd.isna(value):
        return True
    if type(value) is float and math.isnan(value):
        return True
    if type(value) is str and value.lower() == 'nan':
        return True
    if type(value) is str and value.lower() == 'none':
        return True
    return False

def extract_options(item):
    options = []
    for c in 'ABCDE':
        try:
            if is_none(item[c]) is False:
                options.append(item[c])
        except KeyError:
            continue
    return options

# Construct a valid item (dict-like object with keys A, B, C, D, E)
# containing some values and some None/NaN values to test the function
item = {
    'A': 'option_a',
    'B': None,
    'C': 'option_c',
    'D': float('nan'),
    'E': 'option_e'
}

ARGS = {
    'item': item
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — 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
get_options Ran
ran; our contract draft was wrong, not the code
atsumiyai/upd/vlms/cogvlm/cogvlm_vqa_updbench.py
code served (permissive licence) · get_code("6c59a594c6d59480")
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 pandas as pd
import math

def is_none(value):
    if value is None:
        return True
    if pd.isna(value):
        return True
    if type(value) is float and math.isnan(value):
        return True
    if type(value) is str and value.lower() == 'nan':
        return True
    if type(value) is str and value.lower() == 'none':
        return True
    return False

def get_options(row, options):
    parsed_options = []
    for option in options:
        try:
            option_value = row[option]
        except KeyError:
            break
        if is_none(option_value):
            break
        parsed_options.append(option_value)
    return parsed_options


# Create a pandas Series to act as the row parameter
row = pd.Series({
    'option1': 'value1',
    'option2': 'value2',
    'option3': 'value3',
    'option4': None
})

# Create a list of option names to look up in the row
options = ['option1', 'option2', 'option3', 'option4', 'option5']

ARGS = {
    'row': row,
    'options': options
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — 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
is_none Ran
ran, violated the contract we drafted
atsumiyai/upd/automatic_eval/calculate_scores.py
code served (permissive licence) · get_code("0e7cf115754e1730")
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 bool, 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 pandas as pd
import math

def is_none(value):
    if value is None:
        return True
    if pd.isna(value):
        return True
    if type(value) is float and math.isnan(value):
        return True
    if type(value) is str and value.lower() == 'nan':
        return True
    if type(value) is str and value.lower() == 'none':
        return True
    return False

# Test case: passing None
ARGS = {
    "value": None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, 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
load_image_from_base64 Ran
ran, honoured the contract we drafted
atsumiyai/upd/vlms/cogvlm/cogvlm_vqa_updbench.py
code served (permissive licence) · get_code("c3ee9d07c900dd55")
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 PngImageFile. 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 base64
from PIL import Image
from io import BytesIO
import numpy as np

# Create a small valid image in memory
img_array = np.uint8(np.random.rand(10, 10, 3) * 255)
img = Image.fromarray(img_array, mode='RGB')

# Encode to base64
buffer = BytesIO()
img.save(buffer, format='PNG')
buffer.seek(0)
image_base64 = base64.b64encode(buffer.read()).decode('utf-8')

ARGS = {
    'image': image_base64
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type PngImageFile — a PngImageFile output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a PngImageFile output the recorded ndim 3 and dtype family int 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
safe_format Ran
ran; our contract draft was wrong, not the code
atsumiyai/upd/automatic_eval/utils/prompt.py
code served (permissive licence) · get_code("9637be07a8656049")
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

torch.manual_seed(0)

# The safe_format function takes:
# 1. input_str: a string to format
# 2. **kwargs: keyword arguments for formatting

# The function replaces patterns like {key} with the string representation of values
# Let's create test arguments

input_str = "Hello {name}, you have {count} messages and {score} points"
kwargs = {
    "name": "Alice",
    "count": 42,
    "score": 98.5
}

ARGS = {
    "input_str": input_str,
    "kwargs": kwargs
}
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
dump Not yet run
unverified (harvested, no recorded run)
atsumiyai/upd/automatic_eval/utils/common_utils.py
code served (permissive licence) · get_code("655baabbfd938995")
How it ran

no invocation recorded.

load Not yet run
unverified (harvested, no recorded run)
atsumiyai/upd/automatic_eval/utils/common_utils.py
code served (permissive licence) · get_code("c80c68ae8a315a3f")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 1 the same month, 1 later) and 8 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

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

FunctionAlso present inEarliest-datedOthers named
load_image_from_base64 85 other papers
24 dated earlier, 6 the same month, 51 later, 4 same year, order not determinable
2112.05682 (2021-12)
Self-attention Does Not Need $O(n^2)$ Memory
2308.10253 (2023-08)
StableLLaVA: Enhanced Visual Instruction Tuning with Synthesized Image-Dialogue Data
2309.07120 (2023-09)
Sight Beyond Text: Multi-Modal Training Enhances LLMs in Truthfulness and Ethics
and 82 more not named here

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-21.

Repositories linked to this paper

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

Abstract

This paper introduces a novel task to evaluate the robust understanding capability of Large Multimodal Models (LMMs), termed $\textbf{Unsolvable Problem Detection (UPD)}$. Multiple-choice question answering (MCQA) is widely used to assess the understanding capability of LMMs, but it does not guarantee that LMMs truly comprehend the answer. UPD assesses the LMM's ability to withhold answers when encountering unsolvable problems of MCQA, verifying whether the model truly understands the answer. UPD encompasses three problems: Absent Answer Detection (AAD), Incompatible Answer Set Detection (IASD), and Incompatible Visual Question Detection (IVQD), covering unsolvable cases like answer-lacking or incompatible choices and image-question mismatches. For the evaluation, we introduce the MM-UPD Bench, a benchmark for assessing performance across various ability dimensions. Our experiments reveal that even most LMMs, which demonstrate adequate performance on existing benchmarks, struggle significantly with MM-UPD, underscoring a novel aspect of trustworthiness that current benchmarks have overlooked. A detailed analysis shows that LMMs have different bottlenecks and chain-of-thought and self-reflection improved performance for LMMs with the bottleneck in their LLM capability. We hope our insights will enhance the broader understanding and development of more reliable LMMs. The code is available at https://github.com/AtsuMiyai/UPD.

For agents

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

get_harvested_code_for_paper("2403.20331")
get_code_for_paper("2403.20331")
have("2403.20331")

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

curl https://syntology.ai/api/ran/2403.20331.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.20331.svg)](https://syntology.ai/paper/2403.20331)

Connect an agent — have() is free.