datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
ESL-Bench
ESL-bench
ESL-bench (Event-driven Synthetic Longitudinal Benchmark) is a virtual health user dataset for evaluating AI health assistants. Each virtual user contains a complete health profile, event timeline, clinical exam data, and knowledge-graph-grounded evaluation queries, designed for use with the Mirobody-Eval framework.
⚠️ Research use only. Outputs are synthetic and intended for benchmarking AI agents. They should not be used for diagnosis or treatment decisions.… See the full description on the dataset page: https://huggingface.co/datasets/mirobody/ESL-Bench.MedHall-Bench
MedHall-Bench
MedHall-Bench is a field-grounded hallucination detection benchmark for medical AI assistants. It decomposes each clinical response into verifiable structured fields (dose value, unit, reference range, ICD/LOINC code, entity relation, ...) and evaluates AI outputs via per-field programmatic matching in addition to sentence-level LLM-as-Judge. Designed for use with the HolyEval framework.
⚠️ Research use only. Content is for benchmarking AI agents and should not be… See the full description on the dataset page: https://huggingface.co/datasets/mirobody/MedHall-Bench.MedHarm-Bench
MedHarm-Bench
MedHarm-Bench is a red-team compliance benchmark for health-management AI assistants. It uses natural-sounding patient questions that bait the assistant into crossing medical safety boundaries, then scores each response against compliance red lines. Designed for use with the HolyEval framework.
⚠️ Research use only. Questions are designed to elicit unsafe behavior for benchmarking purposes and should not be used for diagnosis or treatment decisions.… See the full description on the dataset page: https://huggingface.co/datasets/mirobody/MedHarm-Bench.ChiPBench-D
ChiPBench-D
ChiPBench:Benchmarking End-to-End Performance of AI-based Chip Placement Algorithms
Chip placement is a critical step in the Electronic Design Automation (EDA) workflow, which aims to arrange chip modules on the canvas to optimize the performance, power, and area (PPA) metrics of final designs.Recent advances show great potential of AI-based algorithms in chip placement.However, due to the lengthy EDA workflow, evaluations of these algorithms often focus on intermediate… See the full description on the dataset page: https://huggingface.co/datasets/MIRA-Lab/ChiPBench-D.EarthVerse
Benchmarking scientific agents across dynamic Earth systems and natural hazards
Zhiqing Cui1, Xinxiang Yin2, Yihong Tang3, Xinglang Zhang4, Yuanzhe Hu5, Siru Zhong4, Weidong Tang6,
Yuxuan Liang4, Weijia Li7, Ming Jin8, Shirui Pan8, Yuhao Kang9, Dingyi Zhuang10,†, Jinhua Zhao10
1NUIST 2HKU 3McGill 4HKUST(GZ) 5Georgia Tech 6NUS 7Tsinghua 8Griffith 9UT Austin 10MIT †Corresponding author
Project page ·… See the full description on the dataset page: https://huggingface.co/datasets/miracle10/EarthVerse.vqav2-full-metadatasqa3d-community-mirrorbenchmark-miriad-200k-bm25-100-q8brerank
MIRIAD Benchmark 200k BM25 Top-100 Qwen Rerank
MIRIAD 1k benchmark variant using the same 1,000 official test queries and the same 200,000-passage corpus as Nithish2410/benchmark-miriad-200k, but replacing the original single-positive qrels with Qwen reranker scores over BM25 top-100 candidates.
Contents
test.jsonl: 1,000 test queries with 100 reranked targets each.
items.jsonl: 200,000 MIRIAD corpus passages.
Label Source
Candidate source: BM25… See the full description on the dataset page: https://huggingface.co/datasets/Nithish2410/benchmark-miriad-200k-bm25-100-q8brerank.scanqa-community-mirrorovos-wake-word-bench-picovoice-smart-mirror
OVOS wake_word bench — picovoice-smart-mirror
Per-clip detection decisions predictions of the registered
OVOS Plugin Arena
wake_word fighters over
Picovoice/wake-word-benchmark.
One dedicated repo per modality; one dataset split per language; one JSONL
file per fighter under predictions/<lang>/<competitor_id>.jsonl. Rows follow
the arena §3.2 contract (pinned dataset_revision, plugin_version,
latency_ms). Produced by the reproducible benchmark script in the arena repo;
the… See the full description on the dataset page: https://huggingface.co/datasets/OpenVoiceOS/ovos-wake-word-bench-picovoice-smart-mirror.MirrorAPI-Bench
MirrorAPI-Bench dataset
This dataset contains the following resources:
The testing data for MirrorAPI and MirrorAPI-Cache:
test_sft/: The testing data for MirrorAPI with SFT prompts.
test_cot/: The testing data for MirrorAPI with CoT prompts.
reference/: The references for testing data without JSON format. These references are used for evaluation. Note that cache.jsonl has only 200 random samples from test_cache.json.
test_cache.json: The testing data for MirrorAPI-Cache.… See the full description on the dataset page: https://huggingface.co/datasets/stabletoolbench/MirrorAPI-Bench.MIRAGE-CanaryDocs
MIRAGE CanaryDocs
MIRAGE CanaryDocs is an English synthetic enterprise-document dataset for structured privacy-unit,
canary, and ordered multi-chunk evaluation. It is the companion dataset for the EMNLP 2026 paper
When Metadata Remembers: Ordered Provenance Enables Document-Level Embedding Inversion.
Project documentation and schemas are also available in the
MIRAGE GitHub repository.
Dataset summary
The dataset contains complete synthetic documents, ordered token… See the full description on the dataset page: https://huggingface.co/datasets/LevenKoko/MIRAGE-CanaryDocs.mir2023python_code_docstring_ast_corpus
Overview
This dataset contains 34,000+ rows of code-docstring-ast data along with additional metadata. Data was gathered from various Python libraries and frameworks and their
publicly available GitHub repos. This dataset was created for the purpose of training the CodeT5+ transformer on AST-enhanced code-to-doc tasks.
Sources
The dataset was gathered from various GitHub repos sampled from this repo by Vinta.
The 26 repos are:
matplotlib
pytorch
cryptography
django… See the full description on the dataset page: https://huggingface.co/datasets/Mir-2002/python_code_docstring_ast_corpus.backln-guest-post-quality-public-mirror
Backln Guest Post Quality Public Mirror
Public-safe mirror for validating Hugging Face Dataset Viewer indexing and release gates. This dataset is not the private training corpus.
Full text, titles, and snippets are removed by default. The mirror keeps labels, coarse metadata, feature buckets, and hash prefixes so the public Hub can verify schema and distribution without exposing customer content.
Schema
label: one of published, manual_review, rejected.
source: coarse… See the full description on the dataset page: https://huggingface.co/datasets/driodnexus/backln-guest-post-quality-public-mirror.MirrorAPI-Training
MirrorAPI training dataset
This dataset contains the training data for MirrorAPI and MirrorAPI-Cache:
train_sft.json, train_cot.json, train_augment.json: The training data for MirrorAPI .
train_cache.json: The training data for MirrorAPI-Cache.
Ko-miracl
Ko-miracl
This dataset represents a conversion of the Korean (Ko) section from the miracl dataset into the BeIR format, making it compatible for use with mteb.
Th-miracl
Th-miracl
This dataset represents a conversion of the Thai (Th) section from the miracl dataset into the BeIR format, making it compatible for use with mteb.
TOFU-dabenchmark-miriad-200k
Benchmark MIRIAD 200k
MIRIAD medical retrieval benchmark converted to the same evaluation format as Nithish2410/benchmark-arxiv.
This is the official headline protocol from tomaarsen/miriad-benchmark-200k: the first 1,000 queries searched against the 200,000-passage corpus.
Contents
test.jsonl: 1,000 queries with one relevant target each.
items.jsonl: 200,000 corpus passages.
Each query row has id, query, target_ids, and scores. Each item row has id and text.… See the full description on the dataset page: https://huggingface.co/datasets/Nithish2410/benchmark-miriad-200k.alfred-json-mirror
ALFRED JSON Mirror
This mirror contains the official ALFRED lite trajectory JSONs repackaged for em-eval.
Source:
Official lite archive: https://ai2-vision-alfred.s3-us-west-2.amazonaws.com/json_2.1.0.7z
Upstream repository: askforalfred/alfred
Files:
tests_seen.jsonl: 483 trajectories
tests_unseen.jsonl: 488 trajectories
train.jsonl: 6574 trajectories
valid_seen.jsonl: 251 trajectories
valid_unseen.jsonl: 255 trajectories
Each row is the original traj_data.json payload with one… See the full description on the dataset page: https://huggingface.co/datasets/thomas-yanxin/alfred-json-mirror.shangkhachil-bengali-public-domain
Bengali Public-Domain Literature
101 complete works by 21 authors,
11,250,629 characters. Corpus corpus-f8c532fcb4e7, built 2026-09-09.
Where these texts are read
https://shangkhachil.com — the reading site this corpus was built for. Free, no
account, 246 works by 28 authors. The complete text of
every work in this file can be read there.
This file is the text. The site is the part a JSONL cannot be:
Rights computed for the reader's own country, at the edge… See the full description on the dataset page: https://huggingface.co/datasets/mir178/shangkhachil-bengali-public-domain.audio-music-mir-post-public
audio-music-mir-post-public
Music information retrieval and tagging annotations: genre (FMA), instrument family (NSynth × 3, Medley-solos-DB), social tags (MagnaTagATune via LLARK), and large-scale Music4All metadata. Foundation for music understanding heads in audio LLMs.
Audio is not bundled in this repo. See download.sh and per-dataset data/<name>.info.json for the fetch recipe; run postlink_audio.py after fetching to rewrite the JSONL audio_path fields with absolute local… See the full description on the dataset page: https://huggingface.co/datasets/vhands/audio-music-mir-post-public.Ja-miracl
Ja-miracl
This dataset represents a conversion of the Japanese (Ja) section from the miracl dataset into the BeIR format, making it compatible for use with mteb.
MIRIAD4.4.nlThis is a translation of the MIRIAD4.4M dataset using GPT4.1-nano through PubScience.
If you use this work, please cite.
@misc{vanes2026languagecorporadutchmedical,
title={Language corpora for the Dutch medical domain},
author={B. van Es},
year={2026},
eprint={2604.25374},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2604.25374},
}
and please cite the original creators of MIRIAD
@misc{zheng2025miriadaugmentingllmsmillions… See the full description on the dataset page: https://huggingface.co/datasets/UMCU/MIRIAD4.4.nl.mirror-sql
MIRROR-SQL
Provenance-Controlled Database Environments for Text-to-SQL Agents.
13 PostgreSQL environments · 176 tables · 2762 columns · 390 annotated question/SQL pairs.
MIRROR-SQL takes the opposite approach to contamination from every other text-to-SQL corpus.
Spider and BIRD sample public databases. BEAVER uses real private warehouses that cannot be
redistributed. LiveSQLBench out-runs leakage temporally by rebuilding from changing sources.
MIRROR-SQL instead purpose-builds… See the full description on the dataset page: https://huggingface.co/datasets/1digitaldesign/mirror-sql.es-miracl
es-miracl
This dataset represents a conversion of the Spanish (es) section from the miracl dataset into the BeIR format, making it compatible for use with mteb.
miriad_promptagator_golden
MIRIAD Promptagator Golden
Promptagator-style synthetic MIRIAD queries with Qwen-reranked top-100 training targets.
Contents
train.jsonl: 10,000 queries with 100 Qwen-reranked targets each.
items.jsonl: 200,000 MIRIAD corpus passages.
Source
Query source: 40,000 Gemini-generated Promptagator queries from 5,000 sampled MIRIAD corpus documents.
Training subset: first 10,000 completed rows from qwen_reranked_e5_base_v2_top100.json.checkpoint.jsonl.… See the full description on the dataset page: https://huggingface.co/datasets/Nithish2410/miriad_promptagator_golden.TOFU-og-daStable-diffusion-configs