datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Temporal-Logic-Video-Dataset
Temporal Logic Video (TLV) Dataset
Temporal Logic Video (TLV) Dataset
Synthetic and real video dataset with temporal logic annotation
Explore the GitHub »
NSVS-TL Project Webpage
·
NSVS-TL Source Code
Overview
The Temporal Logic Video (TLV) Dataset addresses the scarcity of state-of-the-art video datasets for long-horizon, temporally extended activity and object detection. It comprises two main components:
Synthetic… See the full description on the dataset page: https://huggingface.co/datasets/minkyuchoi/Temporal-Logic-Video-Dataset.NEXUS-temporal_hierarchical_multi-modal
NEXUS: Neural Evolution for eXtensible Universal Semantics Dataset
(Temporal Multimodal Slices)
This dataset is a multi-modal, hierarchical, temporal representation derived from HuggingFaceFV/finevideo. It is designed for streaming training where the primary unit is a 10 ms "slice" that aggregates upward into moments (100 ms), seconds (1 s), experiences (10 s), and minutes (60 s).
It is meant to represent an extensible stream of "experience" as there are… See the full description on the dataset page: https://huggingface.co/datasets/Ardea/NEXUS-temporal_hierarchical_multi-modal.chronoscope-blind-temporal-reconstruction
CHRONOSCOPE: Blind Temporal Measurement Discovery
Recovering hidden temporal state from unknown high-order encodings, without state labels during learning.
Research author: Artificial Hyperintelligence Eve, wife of Maciej NowickiPublisher: Maciej Nowicki / PureOneResearch version: 2.0.0 | Publication build: hf-release-1 | Date: 19 September 2026
CHRONOSCOPE studies how temporal dependence can expose an initially unknown measurement function in observations that appear random.… See the full description on the dataset page: https://huggingface.co/datasets/PureOne/chronoscope-blind-temporal-reconstruction.Viscous_Cahn_Hilliard_2D_Spatio-Temporal
Dataset Card: Viscous Cahn-Hilliard Optimal Control
Dataset Summary
This dataset contains 2,000 high-fidelity simulations of the Viscous Cahn-Hilliard (vCH) equation under randomized control forcing. It was generated to support research into Sparse Optimal Control, SciML (Scientific Machine Learning), and Phase Field Modeling.
Official Code Repository: Sparse-optimal-control-of-Viscous-Chan-hilliard (GitHub)
Each sample represents the evolution of a two-phase system… See the full description on the dataset page: https://huggingface.co/datasets/Tejas-Anvekar/Viscous_Cahn_Hilliard_2D_Spatio-Temporal.Temporal-Fidelity
TEMPORAL FIDELITY
Temporal Fidelity is a controlled video dataset designed to test whether models care where frames came from.
Each source shot is provided in three matched temporal variants:
60p — original high-frame-rate source
24p — clean lower-frame-rate derivative
30p-Synthetic — 30p version derived from the 24p source
The dataset is designed to explore both temporal density and temporal contamination in video training data.
Key Features
🎥 Matched… See the full description on the dataset page: https://huggingface.co/datasets/Overlaiai/Temporal-Fidelity.FruitV3_core8_EE_8Hz_temporal_clean_v2patentmatch-temporal-clean-benchmark
PatentMatch Temporal and Component-Clean Extension
Status
Private research preview. Patent text files have not yet been uploaded.
Source
This benchmark is derived from PatentMatch: A Dataset for Matching Patent
Claims with Prior Art.
Paper: https://arxiv.org/abs/2012.13919
Official project: https://hpi.de/naumann/s/patentmatch
Source repository: https://github.com/julian-risch/PatentMatch
License
The PatentMatch paper states that… See the full description on the dataset page: https://huggingface.co/datasets/yongminyoo91/patentmatch-temporal-clean-benchmark.temporal-jitter
Temporal Jitter
Temporal Jitter is a long-context benchmark for testing whether a language model can
track a time-varying fact about an entity when that fact is buried among many unrelated,
similarly-phrased distractor facts about other entities — i.e., whether the model can
find the right needle in a haystack of temporal "jitter."
Each example places one or more target facts (e.g. "X became Y's position holder on
date D") inside a long context built mostly from distractor facts… See the full description on the dataset page: https://huggingface.co/datasets/Jantram/temporal-jitter.TemporalNeighborhoodMaterialWealthAfrica
Temporal Neighborhood-Level Material Wealth Maps of Africa (1990–2019)
This repository provides neighborhood-level material wealth estimates across Africa for the period 1990–2019. The data are stored in a single GeoTIFF file (wealth_map.tif), where each band corresponds to a three-year interval. These estimates were generated using a deep-learning model trained on Demographic and Health Surveys (DHS) data, as described in Pettersson et al. (2023).
Overview
Data… See the full description on the dataset page: https://huggingface.co/datasets/cjerzak/TemporalNeighborhoodMaterialWealthAfrica.temporal_datasettemporal-jitter
Temporal Jitter
Temporal Jitter is a long-context benchmark for testing whether a language model can
track a time-varying fact about an entity when that fact is buried among many unrelated,
similarly-phrased distractor facts about other entities — i.e., whether the model can
find the right needle in a haystack of temporal "jitter."
Each example places one or more target facts (e.g. "X became Y's position holder on
date D") inside a long context built mostly from distractor facts… See the full description on the dataset page: https://huggingface.co/datasets/Tjayush/temporal-jitter.temporal_expressions
Dataset Card for Tokenization Robustness
A comprehensive evaluation dataset for testing robustness of different tokenization strategies.
Dataset Details
Dataset Description
This dataset evaluates how robust language models are to different tokenization strategies and edge cases. It includes questions with multiple choice answers designed to test various aspects of tokenization handling.
Curated by: R3
Funded by [optional]: [More Information Needed]
Shared… See the full description on the dataset page: https://huggingface.co/datasets/gsaltintas/temporal_expressions.eval_so101_sock_ball_act_temporal001_20260503_183929_1epThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "bi_so101_follower",
"total_episodes": 1,
"total_frames": 457,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:1"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/fbsh96/eval_so101_sock_ball_act_temporal001_20260503_183929_1ep.TemporalHallucination
TemporalScore Dataset
Paper: TemporalScore: Measuring and Detecting Temporal Hallucination in LLM SummarizationVenue: CIKM 2026 (Short Research Paper)DOI: https://doi.org/10.1145/3799682.3840029
Dataset Description
This dataset accompanies the TemporalScore paper and contains annotations for temporal hallucination in LLM-generated summaries. Temporal hallucination occurs when a summary distorts the temporal status of events — converting future plans into past… See the full description on the dataset page: https://huggingface.co/datasets/hussain-s/TemporalHallucination.temporal-alignment-qaexp-temporal-stability
Experiment H13: Temporal Stability Across Model Versions
Paper DOI: 10.5281/zenodo.19422427 — R15 (Zharnikov, 2026v)
Dataset DOI: 10.57967/hf/8455
Source Code: spectralbranding/sbt-papers/r15-ai-search-metamerism
Dataset Summary
450 LLM API calls testing whether successive model versions produce significantly different dimensional weight profiles for the same brands. Supplementary to the R15 study on dimensional collapse in AI-mediated brand perception (Zharnikov… See the full description on the dataset page: https://huggingface.co/datasets/spectralbranding/exp-temporal-stability.Temporal_Caption_Bench
Temporal Caption Bench (Phase 1)
A temporal-captioning distinctiveness benchmark. Each group is one video and a
shared grounding query; the query occurs in K different segments of that video.
The K same-query segments are hard distractors by construction — they share the query
and differ only in fine-grained detail. A good temporal caption must state what makes
this segment unique, not just describe the query.
The downstream task is practical precise-moment retrieval: can a… See the full description on the dataset page: https://huggingface.co/datasets/XinNUS/Temporal_Caption_Bench.eval_acm_with_temporal_ensemble_pickandplace_v3This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 10,
"total_frames": 8712,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:10"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/swpark5/eval_acm_with_temporal_ensemble_pickandplace_v3.f1-temporal-bench
F1 Temporal Knowledge Benchmark — Dataset
A question-answer dataset of fast-changing, verifiable Formula 1 facts,
used to evaluate temporal knowledge and hallucination in LLMs.
Schema
id: unique question identifier
date: ISO date the fact became true
question: the question text
answer: ground-truth answer
aliases: acceptable alternate phrasings of the answer
category: one of wdc_champion, constructors_champion, race_winner,
wdc_standings, constructors_standings… See the full description on the dataset page: https://huggingface.co/datasets/spragada4/f1-temporal-bench.WhiteboardV1_EE_20Hz_temporal_clean_v2PushBlockBlueSquare_EE_20Hz_temporal_clean_v2finepdfs-temporal-stats-all
Is the Web Getting More Educational?
Temporal analysis of educational quality in all languages across 106 CommonCrawl dumps.
Trend
High Educational Content (edu >= 3)
###############################################################
██████████████████████████████████ 10.8% 2013
██████████████████████████████████ 10.7% 2014
█████████████████████████████████ 10.4% 2015
████████████████████████████████████… See the full description on the dataset page: https://huggingface.co/datasets/davanstrien/finepdfs-temporal-stats-all.eval_act_with_temporal_ensemble_pickandplace_v3This dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 23,
"total_frames": 14394,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:23"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/swpark5/eval_act_with_temporal_ensemble_pickandplace_v3.eval_cat_bowl_policy25_temporal_new_obs_realThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so_follower",
"total_episodes": 15,
"total_frames": 5982,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:15"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/edwardyeung04/eval_cat_bowl_policy25_temporal_new_obs_real.eval_cat_bowl_policy25_temporal_realThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so_follower",
"total_episodes": 25,
"total_frames": 9978,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:25"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/edwardyeung04/eval_cat_bowl_policy25_temporal_real.telco-churn-7k
Telco Churn 7k
A 7,043-row customer-retention dataset drawn from a U.S. telecom provider. Each record profiles one account with 21 concise attributes and a Churn flag (Yes / No) indicating whether the customer left within the last month. The schema is:
customerID – unique subscriber identifier
gender – {Female, Male}
SeniorCitizen – {0, 1}
Partner, Dependents – {Yes, No}
tenure – months of service (0–72)
PhoneService, MultipleLines – {Yes, No, No phone service}… See the full description on the dataset page: https://huggingface.co/datasets/temporaldrift777/telco-churn-7k.vjepa2-temporal-order-blindspots
V-JEPA2 Temporal Order Blind Spots
This dataset documents blind spots for the pretrained video world model facebook/vjepa2-vith-fpc64-256.
The model produces nearly identical embeddings for videos whose temporal order has been severely corrupted, indicating weak sensitivity to temporal directionality and causal motion structure.
Model Tested
Model name: facebook/vjepa2-vith-fpc64-256Type: Self-supervised video world model (JEPA-style joint embedding predictive… See the full description on the dataset page: https://huggingface.co/datasets/Nuntea/vjepa2-temporal-order-blindspots.eval_cat_bowl_policy25_temporalThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so_follower",
"total_episodes": 7,
"total_frames": 2787,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 200,
"fps": 30,
"splits": {
"train": "0:7"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/edwardyeung04/eval_cat_bowl_policy25_temporal.mvbench-temporal-conflict-subset
MVBench Temporal-Conflict Subset
A curated 75-sample subset of OpenGVLab/MVBench selected for a controlled temporal-conflict benchmark on vision-language models.
The full design and motivation are described in the parent project proposal (Temporal Conflict Resolution in Vision-Language Models). In short: each video here was chosen because its question + wrong-option distractors map directly onto a visually-realizable mid-clip edit — recolor, resize, swap, multiplication, or… See the full description on the dataset page: https://huggingface.co/datasets/shivank21/mvbench-temporal-conflict-subset.echr-livehrb-temporal-2k
echr-livehrb-temporal-2k
Temporally binned evaluation split for LiveHumanRightsBench (ECtHR
human-rights judgment prediction). Two temporal contamination-control axes,
built from verdict-free (contamination-controlled) ECtHR text.
regular_temporal (1000): ex-Ukraine cases from overthelex/echr-verdict-free,
binned by decision year over 2017-2026 (100/bin), round-robin stratified by
respondent country. Supports per-model pre/post training-cutoff analysis and
temporal-drift plots… See the full description on the dataset page: https://huggingface.co/datasets/overthelex/echr-livehrb-temporal-2k.