SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2307.09112 · NeurIPS · 2023

NU-MCC: Multiview Compressive Coding with Neighborhood Decoder and Repulsive UDF

Min Lin, Gim Lee, Stefan Lionar, Xiangyu Xu

arXiv · PDF · Open in the Atlas

Code that ran

11 of 16 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (6 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 1 violated it and 6 have no contract verdict; 3 failures of our instrument; 5 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/2307.09112.json · badge: /api/ran/2307.09112.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
sail-sg/numcc marked official in Papers with Code
Papers with Code archive
9 ran · 5 no recorded run
copy not recorded community
not in the paper's declared repositories
2 ran · 0 no recorded run
FunctionStatusWhere it lives
CrossTransformerBlock_NoFine Ran
ran on a synthesized input · constructed an object; no result was computed
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("380854eb9ca9ccc7")
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 an instance of its own class, CrossTransformerBlock_NoFine: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

# Define small stand-in arguments
dim_inp = 8
dim = 4
nneigh = 5
args = None

# Create the module instance
module = nn.Module()

ARGS = {
    "dim_inp": dim_inp,
    "dim": dim,
    "nneigh": nneigh,
    "args": args
}
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 CrossTransformerBlock_NoFine — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a CrossTransformerBlock_NoFine 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
DecoderAttentionCenters Ran
ran on a synthesized input · constructed an object; no result was computed
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("c1419437fe7c93a0")
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 an instance of its own class, DecoderAttentionCenters: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'dim': 8,
    'num_heads': 2,
    'qkv_bias': False,
    'attn_drop': 0.0,
    'proj_drop': 0.0,
    'args': None
}
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 DecoderAttentionCenters — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DecoderAttentionCenters 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
DecoderBlockCenters Ran
ran on a synthesized input · constructed an object; no result was computed
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("113ee3c359efd24f")
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 an instance of its own class, DecoderBlockCenters: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn
from timm.models.vision_transformer import Mlp, DropPath

torch.manual_seed(0)

ARGS = {
    'dim': 4,
    'num_heads': 2,
    'mlp_ratio': 4.0,
    'qkv_bias': False,
    'drop': 0.0,
    'attn_drop': 0.0,
    'init_values': 1e-5,
    'drop_path': 0.0,
    'act_layer': nn.GELU,
    'norm_layer': nn.LayerNorm,
    'args': None,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
timm, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DecoderBlockCenters — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DecoderBlockCenters 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
DecoderPredictCenters Ran
ran on a synthesized input · constructed an object; no result was computed
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("1d2f8baaab7992c5")
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 an instance of its own class, DecoderPredictCenters: the run constructed the object and did not call it, so no result was computed. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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 torch.nn as nn
import numpy as np
from timm.models.vision_transformer import Mlp

torch.manual_seed(0)

class ArgsStandIn:
    def __init__(self):
        self.n_groups = 200
        self.device = 'cpu'
        self.drop_path = 0.1

ARGS = {
    'img_size': 224,
    'patch_size': 16,
    'embed_dim': 768,
    'decoder_embed_dim': 512,
    'decoder_depth': 8,
    'decoder_num_heads': 16,
    'mlp_ratio': 4.0,
    'norm_layer': nn.LayerNorm,
    'args': ArgsStandIn()
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, timm, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DecoderPredictCenters — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a DecoderPredictCenters 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
ResnetBlockFC Ran
ran on a synthesized input · constructed an object; no result was computed
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("8e12757dead9cd1f")
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 an instance of its own class, ResnetBlockFC: the run constructed the object and did not call it, so no result was computed. 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 torch.nn as nn

torch.manual_seed(0)

# The ResnetBlockFC class is instantiated with size_in, size_out, and size_h parameters.
# Looking at the __init__ method:
# - size_in: required, input dimension (int)
# - size_out: optional, defaults to size_in (int)
# - size_h: optional, defaults to min(size_in, size_out) (int)

# These are all int parameters defining dimensions for fully connected layers.
# Use distinct sizes for dimensions that mean different things.

ARGS = {
    "size_in": 8,
    "size_out": 6,
    "size_h": 5,
}
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 ResnetBlockFC — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a ResnetBlockFC 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
get_2d_sincos_pos_embed Ran
ran, honoured the contract we drafted
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("caca77a0bdc72e75")
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 ndarray, ndim 2, dtype family float, shape (16, 8) 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 numpy as np

# Define arguments for get_2d_sincos_pos_embed
ARGS = {
    "embed_dim": 8,
    "grid_size": 4,
    "cls_token": False,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (16, 8) — shape (16, 8) 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
index_points Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
this paper's copy was not recorded; identical code first harvested from simongiebenhain/air-nets
pointer only; the stored copy get_code serves, from simongiebenhain/air-nets, is pointer only (licence: NONE) · get_code("c7ac4b209ddab5c9")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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 on a synthesized input · constructed an object; no result was computed
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("2fb812fe9013eeb5")
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 an instance of its own class, positional_encoding: the run constructed the object and did not call it, so no result was computed. 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
import math

torch.manual_seed(0)

# The positional_encoding class takes a basis_function parameter
# From the __call__ method, we can see it expects p to be a tensor with shape (batch, ..., dim)
# and it concatenates along dim=2, so p should have at least 3 dimensions

# Create a small input tensor: shape (batch=2, seq=3, features=4)
p = torch.randn(2, 3, 4)

ARGS = {
    'basis_function': 'sin_cos'
}
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 positional_encoding — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a positional_encoding 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_img Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("a10674d38f745127")
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 4, dtype family float, shape (2, 3, 224, 1) 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 torch.nn.functional as F

torch.manual_seed(0)

# Create a small image tensor with shape (batch, channels, height, width)
# Using batch=2, channels=3, height=800, width=4 for distinct dimensions
x = torch.randn(2, 3, 800, 4)

# The function has two parameters: x and res
# res defaults to 224.0
ARGS = {
    'x': x,
    'res': 224.0
}
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 4, dtype family float, shape (2, 3, 224, 1) — shape (2, 3, 224, 1) 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
shrink_points_beyond_threshold Ran
ran; our contract draft was wrong, not the code
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("15f6581a30d2365f")
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, 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 torch

torch.manual_seed(0)

# Create a small tensor of 3D points (batch_size=2, coordinates=3)
xyz = torch.randn(2, 3, dtype=torch.float32)

# Set threshold to a reasonable value
threshold = 0.5

ARGS = {
    "xyz": xyz,
    "threshold": threshold,
}
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, 3) — shape (2, 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
square_distance Ran
ran, violated the contract we drafted
this paper's copy was not recorded; identical code first harvested from simongiebenhain/air-nets
pointer only; the stored copy get_code serves, from simongiebenhain/air-nets, is pointer only (licence: NONE) · get_code("b1776b6ece5471d2")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

CrossTransformerBlock Not yet run
unverified (harvested, no recorded run)
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("7f93eb2fa37c5e41")
How it ran

no invocation recorded.

FeatureAggregator Not yet run
unverified (harvested, no recorded run)
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("2150a4637b729c46")
How it ran

no invocation recorded.

MCCEncoder Not yet run
unverified (harvested, no recorded run)
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("4ec5f3be0f2dfaf0")
How it ran

no invocation recorded.

NUMCC Not yet run
unverified (harvested, no recorded run)
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("4d48613084231492")
How it ran

no invocation recorded.

XYZPosEmbed Not yet run
unverified (harvested, no recorded run)
sail-sg/numcc/src/model/nu_mcc.py
code served (permissive licence) · get_code("a0139963a83c6736")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 1 the same month, 3 later) and 11 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_2d_sincos_pos_embed
brick of this paper's assembly: helper
95 other papers
29 dated earlier, 1 the same month, 64 later, 1 same year, order not determinable · 33 only as an assembly brick
1807.10221 (2018-07)
Unified Perceptual Parsing for Scene Understanding
2105.12723 (2021-05)
Nested Hierarchical Transformer: Towards Accurate, Data-Efficient and Interpretable Visual Understanding
2111.06377 (2021-11)
Masked Autoencoders Are Scalable Vision Learners
and 92 more not named here
square_distance
brick of this paper's assembly: helper
14 other papers
7 dated earlier, 7 later · 1 only as an assembly brick
2110.11860 (2021-10)
AIR-Nets: An Attention-Based Framework for Locally Conditioned Implicit Representations
2203.06604 (2022-03)
Masked Autoencoders for Point Cloud Self-supervised Learning
and 11 more not named here
ResnetBlockFC
brick of this paper's assembly: helper
6 other papers
6 dated earlier · 4 only as an assembly brick
1812.03828 (2018-12)
Occupancy Networks: Learning 3D Reconstruction in Function Space
2103.16341 (2021-03)
Learning Parallel Dense Correspondence from Spatio-Temporal Descriptors for Efficient and Robust 4D Reconstruction
2104.06849 (2021-04)
LEAP: Learning Articulated Occupancy of People
and 3 more not named here
index_points
brick of this paper's assembly: helper
6 other papers
3 dated earlier, 3 later · 3 only as an assembly brick
2110.11860 (2021-10)
AIR-Nets: An Attention-Based Framework for Locally Conditioned Implicit Representations
2206.04511 (2022-06)
Efficient Human Pose Estimation via 3D Event Point Cloud
2302.14348 (2023-02)
Im2Hands: Learning Attentive Implicit Representation of Interacting Two-Hand Shapes
and 3 more not named here
shrink_points_beyond_threshold
brick of this paper's assembly: helper
1 other paper
1 dated earlier · 1 only as an assembly brick
2301.08247 (2023-01)
Multiview Compressive Coding for 3D Reconstruction

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

Remarkable progress has been made in 3D reconstruction from single-view RGB-D inputs. MCC is the current state-of-the-art method in this field, which achieves unprecedented success by combining vision Transformers with large-scale training. However, we identified two key limitations of MCC: 1) The Transformer decoder is inefficient in handling large number of query points; 2) The 3D representation struggles to recover high-fidelity details. In this paper, we propose a new approach called NU-MCC that addresses these limitations. NU-MCC includes two key innovations: a Neighborhood decoder and a Repulsive Unsigned Distance Function (Repulsive UDF). First, our Neighborhood decoder introduces center points as an efficient proxy of input visual features, allowing each query point to only attend to a small neighborhood. This design not only results in much faster inference speed but also enables the exploitation of finer-scale visual features for improved recovery of 3D textures. Second, our Repulsive UDF is a novel alternative to the occupancy field used in MCC, significantly improving the quality of 3D object reconstruction. Compared to standard UDFs that suffer from holes in results, our proposed Repulsive UDF can achieve more complete surface reconstruction. Experimental results demonstrate that NU-MCC is able to learn a strong 3D representation, significantly advancing the state of the art in single-view 3D reconstruction. Particularly, it outperforms MCC by 9.7% in terms of the F1-score on the CO3D-v2 dataset with more than 5× faster running speed. 3D reconstruction from a single-view RGBD input is a fundamental problem in computer vision with applications in robotics [1, 2] and VR/AR [3]. The state-of-the-art approach for this task is MCC [4], which leverages large-scale multi-view images [5] to develop a scalable model for 3D reconstruction from a single RGB-D image. MCC utilizes the depths and 3D point clouds for supervision obtained using the COLMAP framework [6,7]. By combining Vision Transformer [8,9] with large-scale training, MCC can learn a generalizable textured single-view 3D reconstruction model that generalizes to diverse zero-shot settings. However, we have identified two key limitations of MCC that affect its reconstruction quality and model efficiency. First, the Transformer decoder in MCC directly takes in the 3D locations of query points to predict their respective occupancy and color. Due to the quadratic complexity of Transformers, this approach incurs a high computational cost when the number of query points is large, as is often the case for detailed 3D reconstruction. Second, MCC uses the occupancy field as the underlying 3D representation, which hinders the recovery of high-fidelity geometry and texture details. This can be observed in Figure 1, where the reconstruction lacks intricate details.

For agents

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

get_harvested_code_for_paper("2307.09112")
get_code_for_paper("2307.09112")
have("2307.09112")

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

curl https://syntology.ai/api/ran/2307.09112.json

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

[![Syntology run record](https://syntology.ai/api/ran/2307.09112.svg)](https://syntology.ai/paper/2307.09112)

Connect an agent — have() is free.