Ali Farhadi, Rowan Zellers, Yejin Choi, Ximing Lu, Jack Hessel, ♥ Youngjae, Yu Jae, Sung Park, Jize Cao, Visual Reasoning
9 of 14 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, 0 violated it and 7 have no contract verdict; 2 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/2106.02636.json · badge: /api/ran/2106.02636.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.
| Repository | Role | Ran |
|---|---|---|
| rowanz/merlot | community (classifier label) classifier basis not recorded on the link |
9 ran · 5 no recorded run |
| Function | Status | Where it lives |
|---|---|---|
| absolute_distance | Ran ran on a synthesized input |
rowanz/merlot/downstream/sort_story/score_permutations.py code served (permissive licence) · get_code("62bc4892e2b51c64") How it ranRan 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 float64, ndim 0, 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.
|
| bytes_feature | Ran ran; our contract draft was wrong, not the code |
rowanz/merlot/utils/data_utils.py code served (permissive licence) · get_code("d682a9593710857e") How it ranRan 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 Feature. 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.
|
| cls_metric | Ran ran on a synthesized input |
rowanz/merlot/downstream/vcr/modeling.py code served (permissive licence) · get_code("27a244d81c36fbf6") How it ranRan 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 dict. 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.
|
| fixed_padding | Ran ran on a synthesized input |
rowanz/merlot/utils/vision_transformer.py code served (permissive licence) · get_code("4d045734e57cf54a") How it ranRan 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 EagerTensor, ndim 4, dtype family float, shape (2, 6, 5, 5) 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.
|
| get_shape_list | Ran ran on a synthesized input |
rowanz/merlot/utils/model_utils.py code served (permissive licence) · get_code("54e064e79cb0b5fd") How it ranRan 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 int; 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.
|
| int64_feature | Ran ran; our contract draft was wrong, not the code |
rowanz/merlot/utils/data_utils.py code served (permissive licence) · get_code("328d17201d9a0e68") How it ranRan 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 Feature. 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.
|
| int64_list_feature | Ran ran on a synthesized input |
rowanz/merlot/utils/data_utils.py code served (permissive licence) · get_code("f6d735872f4495ed") How it ranRan 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 Feature. 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.
|
| reduce_rms | Ran ran on a synthesized input |
rowanz/merlot/utils/optimization.py pointer only (licence: MIT); the stored copy get_code serves, from rowanz/grover, is pointer only (licence: Apache-2.0) · get_code("2d64405efb57086c") How it ranThe body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from rowanz/grover, and Apache-2.0, as GitHub detects it on that repository, is in the permissive set, but a licence decision keeps the body a pointer: its licence file names an outside party, so it may cover vendored code rather than the repository owner's files (recorded as pointer:licence_names_third_party_corporation). 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. |
| spearman_acc | Ran ran on a synthesized input |
rowanz/merlot/downstream/sort_story/score_permutations.py code served (permissive licence) · get_code("1f05965c2f684e4d") How it ranRan 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 float64, ndim 0, 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.
|
| create_initializer | Not yet run unverified (harvested, no recorded run) |
rowanz/merlot/utils/transformer.py pointer only (licence: MIT); the stored copy get_code serves, from huawei-noah/Pretrained-Language-Model, is pointer only (licence: NONE) · get_code("e3981aa2fadfb2eb") How it ranThe body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from huawei-noah/Pretrained-Language-Model, and GitHub detected no licence in that repository when it was checked. 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. |
| gelu | Not yet run unverified (harvested, no recorded run) |
rowanz/merlot/utils/model_utils.py code served (permissive licence) · get_code("e1cd654e2fac4f34") How it ranno invocation recorded. |
| get_tpu_run_config | Not yet run unverified (harvested, no recorded run) |
rowanz/merlot/utils/neat_config.py code served (permissive licence) · get_code("8772475a675f1e62") How it ranno invocation recorded. |
| layer_norm | Not yet run unverified (harvested, no recorded run) |
rowanz/merlot/utils/model_utils.py code served (permissive licence) · get_code("1fda499c82a8193a") How it ranno invocation recorded. |
| score_permutation | Not yet run unverified (harvested, no recorded run) |
rowanz/merlot/downstream/sort_story/score_permutations.py code served (permissive licence) · get_code("e17cc4aa02dacfbf") How it ranno invocation recorded. |
Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 distinct functions with an identity, 5 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 1 later) and 9 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.
| Function | Also present in | Earliest-dated | Others named |
|---|---|---|---|
| create_initializer | 9 other papers 8 dated earlier, 1 later · 6 only as an assembly brick |
1706.03762 (2017-06) Attention Is All You Need |
1810.04805 (2018-10) BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding 1905.12616 (2019-05) Defending Against Neural Fake News and 6 more not named here |
| bytes_feature | 8 other papers 8 dated earlier |
1409.1556 (2014-09) Very Deep Convolutional Networks for Large-Scale Image Recognition |
1607.08022 (2016-07) Instance Normalization: The Missing Ingredient for Fast Stylization 1609.05158 (2016-09) Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network and 5 more not named here |
| int64_feature | 7 other papers 7 dated earlier |
1607.08022 (2016-07) Instance Normalization: The Missing Ingredient for Fast Stylization |
1611.07810 (2016-11) A dataset and exploration of models for understanding video data through fill-in-the-blank question-answering 1802.05591 (2018-02) Towards End-to-End Lane Detection: an Instance Segmentation Approach and 4 more not named here |
| int64_list_feature | 5 other papers 5 dated earlier |
1607.08022 (2016-07) Instance Normalization: The Missing Ingredient for Fast Stylization |
1803.10562 (2018-03) ELEGANT: Exchanging Latent Encodings with GAN for Transferring Multiple Face Attributes 1812.00442 (2018-12) Deep Cosine Metric Learning for Person Re-Identification and 2 more not named here |
| reduce_rms | 1 other paper 1 dated earlier |
1905.12616 (2019-05) Defending Against Neural Fake News |
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.
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
As humans, we understand events in the visual world contextually, performing multimodal reasoning across time to make inferences about the past, present, and future. We introduce MERLOT, a model that learns multimodal script knowledge by watching millions of YouTube videos with transcribed speech -in an entirely label-free, self-supervised manner. By pretraining with a mix of both framelevel (spatial) and video-level (temporal) objectives, our model not only learns to match images to temporally corresponding words, but also to contextualize what is happening globally over time. As a result, MERLOT exhibits strong out-of-the-box representations of temporal commonsense, and achieves state-ofthe-art performance on 12 different video QA datasets when finetuned. It also transfers well to the world of static images, allowing models to reason about the dynamic context behind visual scenes. On Visual Commonsense Reasoning, MERLOT answers questions correctly with 80.6% accuracy, outperforming stateof-the-art models of similar size by over 3%, even those that make heavy use of auxiliary supervised data (like object bounding boxes). Ablation analyses demonstrate the complementary importance of: 1) training on videos versus static images; 2) scaling the magnitude and diversity of the pretraining video corpus; and 3) using diverse objectives that encourage full-stack multimodal reasoning, from the recognition to cognition level.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2106.02636")
get_code_for_paper("2106.02636")
have("2106.02636")
The run record, dated, one paper per request, free:
curl https://syntology.ai/api/ran/2106.02636.json
A badge for a README (the split and the date, never a ratio):
[](https://syntology.ai/paper/2106.02636)
Connect an agent — have() is free.