SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2412.06014 · ICLR · 2024

Post-hoc Probabilistic Vision-Language Models

Arno Solin, Zeynep Akata, Martin Trapp, Rui Li, Shyamgopal Karthik, Anton Baumann, Marcus Klasson, Santeri Mentu

arXiv · PDF · Open in the Atlas

Code that ran

7 of 13 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 5 failures of our instrument; 6 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/2412.06014.json · badge: /api/ran/2412.06014.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
naver-ai/prolip canonical (classifier label)
classifier basis: classifier_2025
3 ran · 2 no recorded run
AaltoML/BayesVLM marked official in Papers with Code
Papers with Code archive
4 ran · 4 no recorded run
FunctionStatusWhere it lives
_as_optional_tensor Ran
ran, honoured the contract we drafted
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("84a0c40ffe322d12")
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 takes an optional tensor and returns it if not None, else None
# We'll test both cases: with a tensor and with None

# Case 1: with a tensor
tensor_arg = torch.randn(2, 4)

# Case 2: with None
none_arg = None

# We'll use the tensor case for ARGS
ARGS = {
    "tensor": tensor_arg
}
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
_get_output Ran
ran, honoured the contract we drafted
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("c22a08de9aa51a86")
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 1, dtype family float, shape (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 as nn
from collections import namedtuple

torch.manual_seed(0)

# Create a simple namedtuple-like class to test hasattr branch
OutputWithName = namedtuple('OutputWithName', ['loss', 'logits'])

# Test case 1: outputs is an object with the requested attribute
outputs_with_attr = OutputWithName(loss=torch.tensor([0.5]), logits=torch.randn(2, 4))

# Test case 2: outputs is a tuple/list
outputs_tuple = (torch.tensor([0.1]), torch.tensor([0.2]), torch.tensor([0.3]))

# Test case 3: outputs is a simple list
outputs_list = [torch.randn(2, 4), torch.randn(2, 3)]

# We'll use the first test case for ARGS
ARGS = {
    'outputs': outputs_with_attr,
    'name': 'loss',
    'index': 0
}
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 1, dtype family float, shape (1,) — shape (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
_normalize_mean_and_var Ran
ran; our contract draft was wrong, not the code
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("b2ce6787aa1b9885")
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 3, 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 mean and var
# mean: shape (batch=2, dim=4)
# var: shape (batch=2, dim=4)
mean = torch.randn(2, 4)
var = torch.abs(torch.randn(2, 4)) + 0.1  # ensure positive variance

eps = 1e-06

ARGS = {
    "mean": mean,
    "var": var,
    "eps": eps,
}
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 3, 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
_std_from_var Ran
ran; our contract draft was wrong, not the code
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("d07af93e1bec367e")
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 takes a single parameter 'var' which is either None or a torch.Tensor
# It returns None if var is None, otherwise returns sqrt(var)
# We need to test both cases

# Case 1: var is a torch.Tensor
var_tensor = torch.tensor([[1.0, 4.0, 9.0, 16.0], [25.0, 36.0, 49.0, 64.0]])

ARGS = {
    "var": var_tensor
}
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
batched Ran
ran; our contract draft was wrong, not the code
naver-ai/prolip/src/prolip/zero_shot_classifier.py
pointer only (licence: NOASSERTION) · get_code("41856af97765c7ff")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, naver-ai/prolip (NOASSERTION). 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.

get_input_dtype Ran
ran; our contract draft was wrong, not the code
naver-ai/prolip/src/prolip/model.py
pointer only (licence: NOASSERTION) · get_code("4d038cd90610dfa0")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, naver-ai/prolip (NOASSERTION). 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.

get_model_preprocess_cfg Ran
ran; our contract draft was wrong, not the code
naver-ai/prolip/src/prolip/model.py
pointer only (licence: NOASSERTION) · get_code("feaac0348de86b5a")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, naver-ai/prolip (NOASSERTION). 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.

BayesVLMEmbeddingOutput Not yet run
unverified (harvested, no recorded run)
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("cf74575774d01c26")
How it ran

no invocation recorded.

BayesVLMModel Not yet run
unverified (harvested, no recorded run)
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("b4ce6a07560ec5ce")
How it ran

no invocation recorded.

BayesVLMOutput Not yet run
unverified (harvested, no recorded run)
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("5aff72f64fe28e99")
How it ran

no invocation recorded.

_diag_cov Not yet run
unverified (harvested, no recorded run)
AaltoML/BayesVLM/bayesvlm/hf/modeling_bayesvlm_clip.py
code served (permissive licence) · get_code("16e1a732a7be4602")
How it ran

no invocation recorded.

build_zero_shot_classifier_with_stds Not yet run
unverified (harvested, no recorded run)
naver-ai/prolip/src/prolip/zero_shot_classifier.py
pointer only (licence: NOASSERTION) · get_code("90f3675f84d92e26")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, naver-ai/prolip (NOASSERTION). 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. No invocation recorded.

get_cast_dtype Not yet run
unverified (harvested, no recorded run)
naver-ai/prolip/src/prolip/model.py
pointer only (licence: NOASSERTION) · get_code("0529ec3d12d83d1c")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, naver-ai/prolip (NOASSERTION). 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. No invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 2 the same month, 3 later) and 10 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_cast_dtype 66 other papers
55 dated earlier, 2 the same month, 8 later, 1 same year, order not determinable
2103.00020 (2021-03)
Learning Transferable Visual Models From Natural Language Supervision
2210.08402 (2022-10)
LAION-5B: An open large-scale dataset for training next generation image-text models
2212.07143 (2022-12)
Reproducible scaling laws for contrastive language-image learning
and 63 more not named here
get_input_dtype 29 other papers
24 dated earlier, 1 the same month, 4 later
2103.00020 (2021-03)
Learning Transferable Visual Models From Natural Language Supervision
2210.08402 (2022-10)
LAION-5B: An open large-scale dataset for training next generation image-text models
2212.07143 (2022-12)
Reproducible scaling laws for contrastive language-image learning
and 26 more not named here
get_model_preprocess_cfg 2 other papers
1 dated earlier, 1 later
2410.18857 (2024-10)
Probabilistic Language-Image Pre-Training
2503.08048 (2025-03)
LongProLIP: A Probabilistic Vision-Language Model with Long Context Text

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

Vision-language models (VLMs), such as CLIP and SigLIP, have found remarkable success in classification, retrieval, and generative tasks. For this, VLMs deterministically map images and text descriptions to a joint latent space in which their similarity is assessed using the cosine similarity. However, a deterministic mapping of inputs fails to capture uncertainties over concepts arising from domain shifts when used in downstream tasks. In this work, we propose post-hoc uncertainty estimation in VLMs that does not require additional training. Our method leverages a Bayesian posterior approximation over the last layers in VLMs and analytically quantifies uncertainties over cosine similarities. We demonstrate its effectiveness for uncertainty quantification and support set selection in active learning. Compared to baselines, we obtain improved and well-calibrated predictive uncertainties, interpretable uncertainty estimates, and sample-efficient active learning. Our results show promise for safety-critical applications of large-scale models.

For agents

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

get_harvested_code_for_paper("2412.06014")
get_code_for_paper("2412.06014")
have("2412.06014")

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

curl https://syntology.ai/api/ran/2412.06014.json

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

[![Syntology run record](https://syntology.ai/api/ran/2412.06014.svg)](https://syntology.ai/paper/2412.06014)

Connect an agent — have() is free.