SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.01095 · 2026

NarrativeTrack: Evaluating Entity-Centric Reasoning for Narrative Understanding

Hyeonjeong Ha, Bo Feng, Kaixin Ma, Jinjin Ge, Gargi Chakraborty

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 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, 0 violated it and 0 have no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-22; 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/2601.01095.json · badge: /api/ran/2601.01095.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
ageitgey/face_recognition canonical (classifier label)
classifier basis: census
2 ran · 0 no recorded run
FunctionStatusWhere it lives
image_files_in_folder Ran
ran; our contract draft was wrong, not the code
ageitgey/face_recognition/face_recognition/face_detection_cli.py
code served (permissive licence) · get_code("47152c71328b2e94")
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 os
import tempfile
import shutil

# Create a temporary folder with some image files
temp_folder = tempfile.mkdtemp()

# Create sample image files with various extensions
image_files = [
    "photo1.jpg",
    "photo2.jpeg",
    "photo3.png",
    "photo4.JPG",
    "photo5.PNG",
    "document.txt",
    "readme.md"
]

for filename in image_files:
    filepath = os.path.join(temp_folder, filename)
    with open(filepath, 'w') as f:
        f.write("dummy content")

ARGS = {
    "folder": temp_folder
}
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_image_file Ran
ran, honoured the contract we drafted
ageitgey/face_recognition/face_recognition/api.py
code served (permissive licence) · get_code("460f0c275f62beb6")
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 ndarray, ndim 3, dtype family int, shape (3, 4, 3) 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 PIL.Image
import numpy as np
import io

# Create a small test image in memory
img = PIL.Image.new('RGB', (4, 3), color=(255, 0, 0))
img_bytes = io.BytesIO()
img.save(img_bytes, format='PNG')
img_bytes.seek(0)

ARGS = {
    'file': img_bytes,
    'mode': 'RGB'
}
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 ndarray, ndim 3, dtype family int, shape (3, 4, 3) — shape (3, 4, 3) 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 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 2 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.

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

Multimodal large language models (MLLMs) have achieved impressive progress in vision-language reasoning, yet their ability to understand temporally unfolding narratives in videos remains largely underexplored. Narrative understanding requires more than recognizing isolated events: models must maintain coherent representations of who is doing what, when, and where across scene transitions and temporal gaps. We introduce NARRATIVETRACK, the first benchmark to evaluate narrative understanding in MLLMs through fine-grained entity-centric reasoning. Unlike existing benchmarks limited to short clips or coarse scene-level semantics, we decompose videos into constituent entities and evaluate models using a Compositional Reasoning Progression (CRP), a structured framework that progressively increases narrative complexity across three dimensions: entity existence, entity changes, and entity ambiguity. This progression requires models to move beyond local perception to reasoning about entities' temporal persistence, state changes, and fine-grained perceptual disambiguation. To enable scalable benchmark construction, we develop a fully automated entity-centric pipeline that extracts temporally grounded entity representations and provides the foundation for CRP. Evaluations of state-of-the-art MLLMs reveal that existing models struggle to maintain coherent entity representations under visual transitions and temporal dynamics. Opensource general-purpose MLLMs exhibit strong perceptual grounding but weak temporal continuity, while video-specialized MLLMs capture temporal context yet frequently hallucinate entities' contexts. These findings uncover a fundamental trade-off between perceptual grounding and temporal reasoning, indicating that narrative understanding emerges only from their integration. NARRATIVETRACK provides the first systematic framework to diagnose and advance temporally grounded narrative comprehension in MLLMs.

For agents

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

get_harvested_code_for_paper("2601.01095")
get_code_for_paper("2601.01095")
have("2601.01095")

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

curl https://syntology.ai/api/ran/2601.01095.json

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

[![Syntology run record](https://syntology.ai/api/ran/2601.01095.svg)](https://syntology.ai/paper/2601.01095)

Connect an agent — have() is free.