CoolFace
21 shown

datasets

Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.

modalityAllTextImageAudioTabularVideo
taskAllText generationQuestion answeringImage classificationSpeech
sizeAllUnder 1K1K – 10K100K – 1M1M – 10M
formatAllParquetJSONCSVWebDataset
sortTrendingMost likedMost usedRecently updatedNewest
Clear all
01andleb /causalds CausalDS Evaluation-only benchmark. Please do not use this release in training corpora. The repository contains the complete exam presented in the paper, including private ground truth and held-out test labels, as well as the data used for ablations. CausalDS is a benchmark generator for causal reasoning in agentic data-science workflows. Each benchmark instance is a fully synthetically generated scene: a hidden structural causal model (SCM), generated tabular data, and a… See the full description on the dataset page: https://huggingface.co/datasets/andleb/causalds.imageothern<1K0 likes2.7k downloads2mo agoHugging Face02haorentang /causalphys Causal-VL Dataset Causal reasoning VQA dataset with 4 categories × 4 subcategories (3062 questions). Structure Each subcategory contains: annotations/*.json — question, answer, causal graph data/ — images (.jpg/.png) or videos (.mp4) Categories Category Subcategories Perception optics, containability, Scene_Reconstruction, Mechanics_Reasoning Anticipation Collision_Prediction, deformation, Fluid_Flow, Intention_Speculation Intervention… See the full description on the dataset page: https://huggingface.co/datasets/haorentang/causalphys.imagevisual-question-answeringn<1K1 likes1.2k downloads4mo agoHugging Face03CausalLM /Retrievatar Retrievatar Retrievatar is a multimodal dataset designed to enhance the retrieval-augmented generation capabilities of vision-language models, specifically focusing on fictional anime characters and real-world celebrities across various fields. This release represents a subset of 100,000 samples extracted from a significantly larger synthetic image-text corpus. The dataset is being open-sourced to facilitate further research into entity-centric multimodal understanding, with plans… See the full description on the dataset page: https://huggingface.co/datasets/CausalLM/Retrievatar.imageimage-text-to-text100K<n<1M75 likes918 downloads9mo agoHugging Face04CausalVerse /CausalVerse_Image CausalVerse Image Dataset This dataset contains two families of splits: Physics splits: Fall, Refraction, Slope, Spring Static image generation: scene1, scene2, scene3, scene4 All splits share the same columns: image (binary image; datasets.Image) render_path (string; original image filename/path) metavalue (string; per-sample metadata; schema varies by split) Paper: CausalVerse: Benchmarking Causal Representation Learning with Configurable High-Fidelity Simulations Project… See the full description on the dataset page: https://huggingface.co/datasets/CausalVerse/CausalVerse_Image.imageimage-feature-extraction100K<n<1M3 likes843 downloads11mo agoHugging Face05anonymous-causal-plan /Causal_Plan Causal Plan Causal Plan is a unified multimodal dataset release for training and evaluating causal reasoning over visually grounded plans. The repository is organized as one entry point with three clearly separated resources: Causal_Plan/ CausalPlan-1M-QA/ CausalPlan-1M-FourStage-Metadata/ Causal-Plan-Bench/ DATASET_MANIFEST.json verify_alignment.py README.md The QA examples, item-level four-stage metadata, and benchmark package are stored in the same repository so that… See the full description on the dataset page: https://huggingface.co/datasets/anonymous-causal-plan/Causal_Plan.imagevisual-question-answering0 likes752 downloads5mo agoHugging Face06Mwxinnn /CausalSpatial CausalSpatial CausalSpatial is a visual question answering benchmark for evaluating object-centric causal spatial reasoning in vision-language models. Each question presents a 3D-rendered scene and asks the model to reason about physical outcomes — not just what is visible, but what would happen given a specific action or trajectory. Dataset Structure Synthetic subsets (collision, compatibility, occlusion, physics) Field Type Description id string… See the full description on the dataset page: https://huggingface.co/datasets/Mwxinnn/CausalSpatial.imagevisual-question-answering1K<n<10K1 likes161 downloads6mo agoHugging Face07BrainCause /Concept_Targeted_Causal_Images Dataset Card for Concept-Targeted Causal Images Dataset Summary Concept-Targeted Causal Images is a concept-centric image dataset designed for studying causal visual representations in the brain. For each concept, the dataset contains three complementary image types: Positive images that clearly depict the target concept Semantic negatives that are visually or semantically related to the concept, but do not satisfy it Counterfactual edits created by editing… See the full description on the dataset page: https://huggingface.co/datasets/BrainCause/Concept_Targeted_Causal_Images.imageimage-classification100K<n<1M4 likes157 downloads4mo agoHugging Face08MM-Hallu /Causal-HalBench Causal-HalBench Benchmark for evaluating spurious correlation-driven hallucinations in LVLMs. 9,709 QA pairs with counterfactual images across 2,144 unique scenes. Fields Field Description image Input image (original or counterfactual) image_name COCO image identifier question Question about the image type Question type (target/distractor) answer Ground truth answer (yes/no) id Unique QA pair identifier tag Image variant tag… See the full description on the dataset page: https://huggingface.co/datasets/MM-Hallu/Causal-HalBench.imagevisual-question-answering1K<n<10K0 likes96 downloads2mo agoHugging Face09Ryukijano /causal-jepa-icml2026-blog-assetsimagen<1K0 likes92 downloads2mo agoHugging Face10monica-sekoyan /CausalClock MeasureBench Clock Hand Counterfactuals This dataset contains 100 source clock designs and 1550 total images. Each source clock has one base image and one isolated counterfactual for each combination of visible hand and perturbation band (Local, Q1, Q2, Q3, Q4). Sources without a second hand additionally have second-hand counterfactuals for Q1--Q4; those rows introduce a visible second hand and intentionally omit Local. Only the selected hand moves in a counterfactual. The clock… See the full description on the dataset page: https://huggingface.co/datasets/monica-sekoyan/CausalClock.imagevisual-question-answering1K<n<10K0 likes57 downloads5d agoHugging Face11nileshsarkar-ai /causal-dimensionality-sae Causal Dimensionality of Transformer Layers — Research Artifacts Large-file artifact store for two companion papers on the causal dimensionality of transformer layers: ICML 2026 Mechanistic Interpretability Workshop — Causal Dimensionality of Transformer Layers: SAE Encoder Filtering and AtP Recall Collapse NeurIPS 2026 (under review) — Causal Dimensionality of Transformer Representations: Measurement, Scaling, and Layer Structure This HuggingFace dataset repo holds the large… See the full description on the dataset page: https://huggingface.co/datasets/nileshsarkar-ai/causal-dimensionality-sae.imagen<1K0 likes52 downloads4mo agoHugging Face12Gokottaw434 /para_Causal_Reasoningimagen<1K0 likes32 downloads11mo agoHugging Face13jiayuttkx /Causal-LERFimagen<1K0 likes32 downloads1mo agoHugging Face14AsphyXIA /causal-vlm-benchimage0 likes26 downloads4mo agoHugging Face15cxhoang /causal-vqaimage1K<n<10K0 likes24 downloads1y agoHugging Face16textual-causal-reasoning /dagverse-exampleimagen<1K1 likes24 downloads6mo agoHugging Face17robomotic /causality-two-room robomotic/causality-two-room This dataset contains trajectories collected in the swm/GlitchedHueTwoRoom-v1 environment for causal world-model experiments. What is inside glitched_hue_tworoom.h5: HDF5 dataset with trajectories and rendered frames. Includes observations, actions, rewards, episode indexing, variation values, and pixel renderings. How it was generated Collection script: scripts/data/collect_glitched_hue.py Configuration:… See the full description on the dataset page: https://huggingface.co/datasets/robomotic/causality-two-room.imagereinforcement-learningn<1K0 likes23 downloads6mo agoHugging Face18liantian /Causal_Worldimagen<1K0 likes10 downloads1y agoHugging Face19liuzhongyan /CausalDiff-Urban-Dataimagen<1K0 likes10 downloads4mo agoHugging Face20Epzq96 /CausalChaos_Features Content Files includes appearance and motion features extracted using resnet 101 and resnext101. h5 files contains id and features under the keys 'ids' and 'feat' respectively. Features are in the format [16, 4096] [:, :2048] contains the appearance features from resnet 101 [:, 2048:] contains the motion features from resnet 101 image0 likes4 downloads2y agoHugging Face21anonymous-submission1012 /CausalConflictBenchimagen<1K0 likes2 downloads5mo agoHugging Face

Listings come live from the Hugging Face Hub API. CoolFace does not host these files.