SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2409.14713 · 2024

Phantom of Latent for Large Language and Vision Models

arXiv · PDF · Open in the Atlas

Code that ran

3 of 4 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, 1 violated it and 0 have no contract verdict; 2 failures of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2409.14713.json · badge: /api/ran/2409.14713.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
byungkwanlee/phantom canonical (classifier label)
classifier basis: census
3 ran · 1 no recorded run
FunctionStatusWhere it lives
get_most_similar Ran
ran; our contract draft was wrong, not the code
byungkwanlee/phantom/eval/mathvista/calculate_score.py
code served (permissive licence) · get_code("c28eb86a8d14e4ed")
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.

from Levenshtein import distance

def get_most_similar(prediction, choices):
    """
    Use the Levenshtein distance (or edit distance) to determine which of the choices is most similar to the given prediction
    """
    distances = [distance(prediction, choice) for choice in choices]
    ind = distances.index(min(distances))
    return choices[ind]


ARGS = {
    "prediction": "kitten",
    "choices": ["sitting", "kitten", "kitchen", "mitten"]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
Levenshtein — 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
normalize_extracted_answer Ran
ran; our contract draft was wrong, not the code
byungkwanlee/phantom/eval/mathvista/calculate_score.py
code served (permissive licence) · get_code("8b901d87160993c2")
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
from Levenshtein import distance

def get_most_similar(prediction, choices):
    """
    Use the Levenshtein distance (or edit distance) to determine which of the choices is most similar to the given prediction
    """
    distances = [distance(prediction, choice) for choice in choices]
    ind = distances.index(min(distances))
    return choices[ind]

def normalize_extracted_answer(extraction, choices, question_type, answer_type, precision):
    """
    Normalize the extracted answer to match the answer type
    """
    if question_type == 'multi_choice':
        if isinstance(extraction, str):
            extraction = extraction.strip()
        else:
            try:
                extraction = str(extraction)
            except:
                extraction = ''
        letter = re.findall('\\(([a-zA-Z])\\)', extraction)
        if len(letter) > 0:
            extraction = letter[0].upper()
        options = [chr(ord('A') + i) for i in range(len(choices))]
        if extraction in options:
            ind = options.index(extraction)
            extraction = choices[ind]
        else:
            extraction = get_most_similar(extraction, choices)
        assert extraction in choices
    elif answer_type == 'integer':
        try:
            extraction = str(int(float(extraction)))
        except:
            extraction = None
    elif answer_type == 'float':
        try:
            extraction = str(round(float(extraction), int(precision)))
        except:
            extraction = None
    elif answer_type == 'list':
        try:
            extraction = str(extraction)
        except:
            extraction = None
    return extraction

ARGS = {
    'extraction': '(B) option two',
    'choices': ['option one', 'option two', 'option three'],
    'question_type': 'multi_choice',
    'answer_type': 'string',
    'precision': 2
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
Levenshtein — 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
safe_equal Ran
ran, violated the contract we drafted
byungkwanlee/phantom/eval/mathvista/calculate_score.py
code served (permissive licence) · get_code("e7a23ea3695f6437")
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 torch
import numpy as np

torch.manual_seed(0)

# The function safe_equal simply compares two values for equality
# It handles exceptions that might arise from the comparison
# We need to provide two arguments: prediction and answer

# Test case 1: Simple scalar values
prediction_1 = 42
answer_1 = 42

# Test case 2: Different types but comparable
prediction_2 = torch.tensor([1.0, 2.0, 3.0])
answer_2 = torch.tensor([1.0, 2.0, 3.0])

# Test case 3: Strings
prediction_3 = "hello"
answer_3 = "hello"

# Test case 4: Lists
prediction_4 = [1, 2, 3]
answer_4 = [1, 2, 3]

# Use the first test case as the primary arguments
ARGS = {
    "prediction": prediction_1,
    "answer": answer_1
}
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 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_model Not yet run
unverified (harvested, no recorded run)
byungkwanlee/phantom/model/load_model.py
code served (permissive licence) · get_code("db11a804a6adefad")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 3 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 2 the same month, 3 later) and 1 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
get_most_similar 7 other papers
3 dated earlier, 1 the same month, 3 later
2403.03003 (2024-03)
Feast Your Eyes: Mixture-of-Resolution Adaptation for Multimodal Large Language Models
2405.05949 (2024-05)
CuMo: Scaling Multimodal LLM with Co-Upcycled Mixture-of-Experts
2405.15574 (2024-05)
Meteor: Mamba-based Traversal of Rationale for Large Language and Vision Models
and 4 more not named here
safe_equal 6 other papers
2 dated earlier, 1 the same month, 3 later
2403.03003 (2024-03)
Feast Your Eyes: Mixture-of-Resolution Adaptation for Multimodal Large Language Models
2405.15574 (2024-05)
Meteor: Mamba-based Traversal of Rationale for Large Language and Vision Models
2409.00147 (2024-09)
MultiMath: Bridging Visual and Mathematical Reasoning for Large Language Models
and 3 more not named here
normalize_extracted_answer 1 other paper
1 later
2509.19003 (2025-09)
Unveiling Chain of Step Reasoning for Vision-Language Models with Fine-grained Rewards

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 success of visual instruction tuning has accelerated the development of large language and vision models (LLVMs). Following the scaling laws of instruction-tuned large language models (LLMs), LLVMs either have further increased their sizes, reaching 26B, 34B, and even 80B parameters. While this increase in model size has yielded significant performance gains, it demands substantially more hardware resources for both training and inference. Consequently, there naturally exists a strong need for efficient LLVMs that achieve the performance of larger models while being smaller in size. To achieve this need, we present a new efficient LLVM family with model sizes of 0.5B, 1.8B, 3.8B, and 7B parameters, Phantom, which significantly enhances learning capabilities within limited structures. By temporarily increasing the latent hidden dimension during multi-head self-attention (MHSA), we make LLVMs prepare to look and understand much more vision-language knowledge on the latent, without substantially increasing physical model sizes. To maximize its advantage, we introduce Phantom Optimization (PO) using both autoregressive supervised fine-tuning (SFT) and direct preference optimization (DPO)-like concept, which effectively follows correct answers while eliminating incorrect and ambiguous ones. Phantom outperforms numerous larger open- and closed-source LLVMs, positioning itself as a leading solution in the landscape of efficient LLVMs.

For agents

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

get_harvested_code_for_paper("2409.14713")
get_code_for_paper("2409.14713")
have("2409.14713")

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

curl https://syntology.ai/api/ran/2409.14713.json

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

[![Syntology run record](https://syntology.ai/api/ran/2409.14713.svg)](https://syntology.ai/paper/2409.14713)

Connect an agent — have() is free.