datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
FINSABER-reproduce
FINSABER Data
Aggregated datasets for the FINSABER backtesting framework (KDD 2026).
Files
File
Description
Size
data/finmem_data/stock_data_sp500_2000_2024.pkl
S&P500 full aggregated data (Price + News + Filings)
~11 GB
data/finmem_data/stock_data_cherrypick_2000_2024.pkl
Selected symbols (TSLA, AMZN, MSFT, NFLX, COIN)
~53 MB
data/price/all_sp500_prices_2000_2024_delisted_include.csv
CSV price-only data for S&P500 (including delisted)
~253 MB… See the full description on the dataset page: https://huggingface.co/datasets/finsaber-team/FINSABER-reproduce.repro-chain-of-thought-gradient-descent-runs-sol
Chain-of-Thought Gradient Descent reproduction runs
Immutable outputs for the independent scaled reproduction of ICML 2026 paper
#443, OpenReview uZ8JZ1Lw9a.
gpu-l4-seed443/: successful NVIDIA L4 checkpoint, result JSON, and cost CSV.
figures/: interactive logbook figures and raw CSVs.
poster/: Posterly source, zero-warning gate report, PDF/PNG, and
self-contained poster_embed.html.
reproduction-bundle/: complete clean download-and-rerun bundle.
Successful Job:… See the full description on the dataset page: https://huggingface.co/datasets/JG1310/repro-chain-of-thought-gradient-descent-runs-sol.repro
RePro: A Benchmark Dataset for Opinion Mining for Brazilian Portuguese
RePro, which stands for "REview of PROducts," is a benchmark dataset for opinion mining in Brazilian Portuguese. It consists of 10,000 humanly annotated e-commerce product reviews, each labeled with sentiment and topic information. The dataset was created based on data from one of the largest Brazilian e-commerce platforms, which produced the B2W-Reviews01 dataset… See the full description on the dataset page: https://huggingface.co/datasets/lucasnil/repro.icml26-repro-nonlinear-autoencoder-pca
Nonlinear autoencoder / PCA reproduction
This bundle reproduces the paper with the authors' pinned release
(SPOC-group/advantage_nonlinearity@4378017) plus an independent direct
population-gradient-flow audit.
run_official_amp_campaign.py: 96 finite-dimensional runs of the released
two-spike AMP at d=256/512 and eight sample ratios.
run_official_ae_campaign.py: 48 full-batch Adam runs of the released tied,
one-neuron ReLU/ELU autoencoder at d=512/1024, with exact PCA and 20,000… See the full description on the dataset page: https://huggingface.co/datasets/SabaPivot/icml26-repro-nonlinear-autoencoder-pca.swe-rebench-repro-artifactsniaf-reproductionpro-familia-intent-dataset-reproductive-health-de
pro familia Intent Dataset – Reproductive Health (DE)
This is a dataset for intent classification in German chat- and voicebot interactions in the domain of reproductive health counselling.
It is a csv-export of the training and test data of the Rasa-based chat- and voicebots developed for pro familia.
The dataset features 88 intents with in total 3269 examples for training and 256 examples for testing.
Uses
This dataset is to be used for research and development of… See the full description on the dataset page: https://huggingface.co/datasets/pro-familia-Bundesverband/pro-familia-intent-dataset-reproductive-health-de.rag-eval-ja-repro
RAG Eval JA Repro
Current version / 現行版: v1.1
2026-07-12 更新（v1.1）: rag_evaluation_master.csv、採用PDF manifest、PDF checksumを更新し、6月30日公開時のローカル精度検証を同じ4条件で再実行しました。旧版の記述は取り消し線で残し、v1.1の値を併記します。
TL;DR (EN): A derived reproducibility dataset for allganize/RAG-Evaluation-Dataset-JA.
It adds (1) derived *_new answer/question columns (with per-item rationale), and (2) a Wayback-pinned + SHA-256 corpus manifest so anyone can fetch byte-identical source PDFs.
The original CSV is not modified;… See the full description on the dataset page: https://huggingface.co/datasets/SakataConsul/rag-eval-ja-repro.vulfix_reproducesyntra-reproduction
Syntra — Independent Blinded Reproduction
An independent, blinded reproduction of the central empirical claim of Syntra: A
Framework for Wiser and More Human AI (H. Axelsson) — that routing a prompt through an
ethics-first orchestration layer (Valon → Modi → Drift) makes a language model reason
more "wisely" than the same model answering directly. We reproduce the paper's own
three-judge × five-dimension LLM-scoring protocol twice: once with the paper-era judge
models, and once… See the full description on the dataset page: https://huggingface.co/datasets/Infektyd/syntra-reproduction.mantis-repro-databes-repro-resultsrepro-knowing-when-to-quit-bundle
Reproduction bundle
This bundle contains the executable local-GPU scaled reproduction, the official metric helper, generated metrics, and the poster artifact.
Sources
Paper: https://huggingface.co/papers/2604.18419
Official code: https://github.com/HenDav/mid-gen-abs
Official trajectory dataset: https://huggingface.co/datasets/HenDav/mid-gen-abs-data
Direct model: https://huggingface.co/microsoft/Phi-3-mini-4k-instruct
Transfer dataset:… See the full description on the dataset page: https://huggingface.co/datasets/JAEMP/repro-knowing-when-to-quit-bundle.2021-final-assisted-reproductive-technology-art-su
2021 Final Assisted Reproductive Technology (ART) Success Rates
Description
Data were updated on September 11, 2024.
ART data are made available as part of the National ART Surveillance System (NASS) that collects success rates, services, profiles and annual summary data from fertility clinics across the U.S. There are four datasets available: ART Services and Profiles, ART Patient and Cycle Characteristics, ART Success Rates, and ART Summary. All four datasets may be… See the full description on the dataset page: https://huggingface.co/datasets/HHS-Official/2021-final-assisted-reproductive-technology-art-su.trace-icml2026-repro-artifactsbp-gaussian-repro-bundle
BP Gaussian Convergence Reproduction Bundle
This directory contains a clean-room numerical audit for Belief Propagation Converges to Gaussian Distributions in Sparsely-Connected Factor Graphs.
Rerun
python -m pip install -r requirements.txt
python reproduce.py --output outputs/audit
The same script was also run as a Hugging Face CPU Job:
https://huggingface.co/jobs/Srishti280992/6a5d4cc1bee6ee1cf4ed1984.
Contents
reproduce.py: fixed-seed… See the full description on the dataset page: https://huggingface.co/datasets/Srishti280992/bp-gaussian-repro-bundle.2020-final-assisted-reproductive-technology-art-su
2020 Final Assisted Reproductive Technology (ART) Success Rates
Description
ART data are made available as part of the National ART Surveillance System (NASS) that collects success rates, services, profiles and annual summary data from fertility clinics across the U.S. There are four datasets available: ART Services and Profiles, ART Patient and Cycle Characteristics, ART Success Rates, and ART Summary. All four datasets may be linked by “ClinicID.” ClinicID is a unique… See the full description on the dataset page: https://huggingface.co/datasets/HHS-Official/2020-final-assisted-reproductive-technology-art-su.2020-final-assisted-reproductive-technology-art-se
2020 Final Assisted Reproductive Technology (ART) Services and Profiles
Description
ART data are made available as part of the National ART Surveillance System (NASS) that collects success rates, services, profiles and annual summary data from fertility clinics across the U.S. There are four datasets available: ART Services and Profiles, ART Patient and Cycle Characteristics, ART Success Rates, and ART Summary. All four datasets may be linked by “ClinicID.” ClinicID is a… See the full description on the dataset page: https://huggingface.co/datasets/HHS-Official/2020-final-assisted-reproductive-technology-art-se.reproscreener_manual_evaluationsbioprobench-repro-artifactsclass-label-string-token-repro
class-label-string-token-repro
Minimal repro dataset for class-label columns whose source CSV token is a label name instead of an integer id.
vulfix_reproduce3vulfix_reproduce2vulfix_reproduce42021-final-assisted-reproductive-technology-art-se
2021 Final Assisted Reproductive Technology (ART) Services and Profiles
Description
ART data are made available as part of the National ART Surveillance System (NASS) that collects success rates, services, profiles and annual summary data from fertility clinics across the U.S. There are four datasets available: ART Services and Profiles, ART Patient and Cycle Characteristics, ART Success Rates, and ART Summary. All four datasets may be linked by “ClinicID.” ClinicID is a… See the full description on the dataset page: https://huggingface.co/datasets/HHS-Official/2021-final-assisted-reproductive-technology-art-se.upiterbarg-lintseq-reproduction2020-final-assisted-reproductive-technology-art-pa
2020 Final Assisted Reproductive Technology (ART) Patient and Cycle Characteristics
Description
ART data are made available as part of the National ART Surveillance System (NASS) that collects success rates, services, profiles and annual summary data from fertility clinics across the U.S. There are four datasets available: ART Services and Profiles, ART Patient and Cycle Characteristics, ART Success Rates, and ART Summary. All four datasets may be linked by “ClinicID.”… See the full description on the dataset page: https://huggingface.co/datasets/HHS-Official/2020-final-assisted-reproductive-technology-art-pa.