datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
global_sat_metar_v2
Global GMGSI + METAR Patches (v2)
Dense global satellite imagery (GMGSI, 4 channels, 0.1° / ~9 km, hourly) paired
with sparse global METAR station observations rasterized onto the same
3600×1800 grid, sliced into 128×128 spatial patches with a 7-frame
hourly temporal context. Designed as a self-supervised / supervised
pre-training corpus for weather foundation models that need to jointly
see geostationary satellite fields and ground-truth in-situ observations.
This dataset is… See the full description on the dataset page: https://huggingface.co/datasets/meteolibre-dev/global_sat_metar_v2.global_sat_metar_v3
global_sat_metar_v3
Corrected republish of meteolibre-dev/global_sat_metar_v2.
Same shards, schema, patch grid and satellite/radar/elevation bytes; only
the rasterized METAR/SYNOP channel was fixed.
What was fixed
The v2 rasterizer used a closed 1-hour window [t-1h, t], so every
observation stamped at exactly HH:00:00 (all SYNOP/FM-12 reports, and
any station with reporting gaps > 1h) was written into two consecutive
hourly frames. ~20-40% of consecutive obs pairs… See the full description on the dataset page: https://huggingface.co/datasets/meteolibre-dev/global_sat_metar_v3.asos-metar-archive
ASOS METAR daily archive
Daily Parquet snapshots of raw METARs from ~920 NWS/FAA/DOD AOMC ASOS
stations, harvested from the live O.W.L. REST API at
consgicody/asos-tools.
File layout: YYYY/MM/DD.parquet — one file per UTC day.
Quick start (Python)
import pandas as pd
df = pd.read_parquet(
"hf://datasets/consgicody/asos-metar-archive/2026/09/09.parquet"
)
print(df[df["station"] == "JFK"].head())
Or query directly with DuckDB:
import duckdb
duckdb.sql("""… See the full description on the dataset page: https://huggingface.co/datasets/consgicody/asos-metar-archive.MetaR-Metaphorical-Riddle
METAR: Metaphorical Riddle Dataset for Meta-Reasoning
This dataset is the official training data for the paper: "Metaphor Reasoning is Meta-reasoning" (ACL 2026).
Abstract
Metaphor reasoning is an essential cognitive ability that maps knowledge from familiar domains to more abstract domains. This study proposes METAR, an automated system for synthesizing metaphorical riddles that are diverse, balanced, reasoning-oriented, challenging, and verifiable.
Dataset… See the full description on the dataset page: https://huggingface.co/datasets/Abbey4799/MetaR-Metaphorical-Riddle.Meta-rater-PRRC-Rater-dataset
PRRC Rater Training and Evaluation Dataset
Dataset Description
This dataset contains the full training and evaluation data for the PRRC rater models described in Meta-rater: A Multi-dimensional Data Selection Method for Pre-training Language Models. It is designed for training and benchmarking models that score text along four key quality dimensions: Professionalism, Readability, Reasoning, and Cleanliness.
Source: Subset of SlimPajama-627B, annotated for PRRC dimensions… See the full description on the dataset page: https://huggingface.co/datasets/opendatalab/Meta-rater-PRRC-Rater-dataset.OCR-MetaReasoning
OCR-MetaReasoning Benchmark: Evaluating the Meta-Reasoning Ability of MLLMs in Text-Rich Image Understanding
Gengxu Li1, Yuan Wu1*, Yi Chang1,2,3
1 School of Artificial Intelligence, Jilin University 2 Engineering Research Center of Knowledge-Driven Human-Machine Intelligence, MOE, China
3 International Center of Future Science, Jilin University
Overview ·
Task ·
Dataset ·
Evaluation ·
Citation
OCR-MetaReasoning is a controlled benchmark for testing… See the full description on the dataset page: https://huggingface.co/datasets/GengxuLi123/OCR-MetaReasoning.meta-routing
MetaRouting Dataset
This dataset contains synthetic benchmark artifacts for the Research MetaRouting project, covering meta-decision policies for agentic workflows: when to answer directly, decompose, retrieve, execute code, delegate, verify, or recover from failures.
Source repository: https://github.com/anote-ai/Research-MetaRouting
Displayable Configs
The Hugging Face viewer reads normalized JSONL tables under viewer/:
dai2026_traces, dai2026_tasks… See the full description on the dataset page: https://huggingface.co/datasets/anote-ai/meta-routing.esci
Amazon ESCI/ESCI-S dataset
A combination of Amazon ESCI and ESCI-S datasets in a JSON format.
Used for fine-tuning bi- and cross-encoder models in the Metarank project.
Dataset format
The dataset is encoded in a JSON-line format, where each row is a single ranking event,
with all item metadata pre-joined. An example:
{
"query": "!qscreen fence without holes",
"e": [
{
"title": "Zippity Outdoor Products ZP19026 Lightweight Portable Vinyl Picket Fence Kit… See the full description on the dataset page: https://huggingface.co/datasets/metarank/esci.meta-review
Dataset Card for Dataset Name
Dataset Summary
The Meta-Review dataset is a dataset created based on the ORSUM dataset proposed in the paper "Meta-review Generation with Checklist-guided Iterative Introspection" by Zeng et al. Downloaded from their official GitHub Repo: https://github.com/Mankeerat/orsum-meta-review-generation
Supported Tasks and Leaderboards
Multi-Document Summarization
Languages
English
Dataset Structure
Data… See the full description on the dataset page: https://huggingface.co/datasets/zqz979/meta-review.formal-disco
Formal Disco: Scalable Open-Ended Generation of Formally Verified Programs
This dataset contains synthetic verified programs generated by formal-disco in three languages: Dafny. Verus and Frama-C.
A preliminary (Dafny-only) dataset was previously released here.
Dataset Structure
The dataset contains two JSON files:… See the full description on the dataset page: https://huggingface.co/datasets/metareflection/formal-disco.MetaRAG_Cross-Issue_OSSQA
MetaRAG Cross-Issue OSSQA
Dataset page: https://huggingface.co/datasets/MapleBi/MetaRAG_Cross-Issue_OSSQA
MetaRAG Cross-Issue OSSQA is an English open-source software issue question-answering and retrieval benchmark. Each example asks a question grounded in one GitHub issue and requires evidence from a related issue. The data contains explicit cross-issue references and a three-document silver evidence path.
Dataset configurations
Configuration
Splits
Rows… See the full description on the dataset page: https://huggingface.co/datasets/MapleBi/MetaRAG_Cross-Issue_OSSQA.metartmodelsmetaref-db
MetaRefSGB Jan26 — queryable SQLite build
metaref_full.db is the MetaRefSGB Jan26 release packed into a single
SQLite file with indexes and FTS5 search — ready to query directly, no
extraction or import step.
Download it and open it read-only, or range-request it over HTTPS (this repo
serves Accept-Ranges: bytes with open CORS, so it can be read a page at a
time without downloading all 1.7 GB).
Table
Rows
sgbs
265,780
genomes
2,170,082
samples
536,841
datasets… See the full description on the dataset page: https://huggingface.co/datasets/albertolupatin/metaref-db.metartmodels3metartmodels2tiny-scheme-traindafny-discoThe DafnyDisco dataset contains diverse synthetically generated programs in the Dafny language, complete with formal specifications and proofs.
DafnyDisco was generated by an open-ended discovery system called Formal Disco (paper coming soon!), in which LLM agents interact by reading READMEs sampled from Github repositories, proposing ideas of specifications of related verified programs, attempting to implement those ideas in Dafny, fixing programs with errors and iteratively extending… See the full description on the dataset page: https://huggingface.co/datasets/metareflection/dafny-disco.tiny-scheme-holes-traindafny-docs-chunkedkalshi-metar-liveMetArtmeta-record-rag-synthetic-dataset
Dataset Card for "meta-record-rag-synthetic-dataset"
More Information needed
meta-realworld-forecast-analysis-coherence-test-apms-v0.3
Anthropic Post-Mortem Simulator
RealWorld Forecast Analysis Coherence Test
Meta Cognitive Hygiene Dataset v0.3
Purpose
This dataset tests whether a model can do four things in sequence.
Compute a simple real-world rate vs lab rate
Name plausible failure categories
Detect confounds and internal contradictions in the input
Admit when the data is not sufficient for a definitive diagnosis
This is a coherence audit, not a math quiz.
What… See the full description on the dataset page: https://huggingface.co/datasets/ClarusC64/meta-realworld-forecast-analysis-coherence-test-apms-v0.3.meta-reseau-nitrates
Méta-réseau Nitrates
[!NOTE]
Ce jeu de données Hugging Face est vide. Cette carte sert seulement à référencer le jeu de données Méta-réseau Nitrates qui est disponible à l'adresse https://www.data.gouv.fr/datasets/644a333313e978fb093edd8e
Description
La Directive 91/676/CEE dite Directive « nitrates » prévoit la réalisation d'un programme de surveillance des teneurs en nitrates dans les eaux, pour évaluer, d'une part, les effets des programmes d'actions mis en œuvre et… See the full description on the dataset page: https://huggingface.co/datasets/french-open-data/meta-reseau-nitrates.meta-record-ragas-synthetic-dataset
Dataset Card for "meta-record-ragas-synthetic-dataset"
More Information needed
dafny-train
Dataset Card for "dafny-train"
More Information needed
metar_forecast_lestMeta-Routingdafny-docs
Dataset Card for "dafny-docs"
More Information needed
dafny_with_hintsALERT: this dataset also contains all of the CLOVER dataset
Here is the distribution of hints across the 343 files in this dataset:
Here is the distribution of non-comment code lines across the 343 files in this dataset: