datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
LeMat-Bulk-MLIP-Hull
LeMat-Bulk MLIP Hull Reference Datasets
This dataset contains materials close to the convex hull computed using various ML interatomic potentials (MLIPs).
Dataset Splits
all: Contains ALL materials with hull energies for all MLIPs (no threshold filtering)
dft, orb, uma, mace_mp, mace_omat: Materials within 0.001 eV/atom of respective hulls
Energy Types
dft: DFT reference energies
orb: ORB model energies
uma: UMA model energies
mace_mp: MACE-MP model energies… See the full description on the dataset page: https://huggingface.co/datasets/LeMaterial/LeMat-Bulk-MLIP-Hull.LEMUR
EU Law Dataset – Category 15.10: Environment
This dataset contains official legal documents from the European Union, collected from the EUR-Lex website, specifically under category 15.10: "Environment". The documents span from the year 1961 to 2025 and are provided in multiple European "languages. The original documents are in PDF format and have been converted into various text-based formats using OLMCR.
The dataset splits represent the different "languages available for each… See the full description on the dataset page: https://huggingface.co/datasets/G4KMU/LEMUR.LEMAS-Dataset-train
Overview
This dataset is part of LEMAS-Project (lemas-project.github.io/LEMAS-Project).
It contains a large-scale training set (150k+ hours) and a curated evaluation set
(500 utterances per language) covering 10 languages, all with word-level alignment.
Fields
key: unique utterance identifier; the first two characters indicate the language ID
audio: relative path to the MP3 audio file (in the eval set, this key is renamed to "file_name" for compatibility with the viewer)… See the full description on the dataset page: https://huggingface.co/datasets/LEMAS-Project/LEMAS-Dataset-train.LeMat-BulkMotivation: check out the blog post https://huggingface.co/blog/lematerial to hear more about the motivation behind the creation of this dataset.
Changelog:
2025.04.17 (hash: NOT YET RELEASED):
We have changed the Yb default pseudopotential to Yb_3 from VASP, this is the same that Materials Project uses. In the previous version we had kept it as Yb, and Materials Project had to Yb-containing materials. Alexandria and OQMD uses Yb. As a result no Yb-containing materials are… See the full description on the dataset page: https://huggingface.co/datasets/LeMaterial/LeMat-Bulk.s2orc_small
Dataset Card for "s2orc_small"
A small split of the s2orc dataset, includes ~900k english papers with abstract included.
See all detailes in the original dataset card - https://huggingface.co/datasets/allenai/s2orc
LeMat-TrajNote: For PBE we are in the process of providing a precomputed energy_corrected scheme based on Materials Project 2020 Compatibility Scheme
Motivation: check out the blog post https://huggingface.co/blog/lematerial to hear more about the motivation behind the creation of our datasets.
Download and use within Python
from datasets import load_dataset
dataset = load_dataset('LeMaterial/LeMat-Traj', 'compatible_pbe')
Data fields
Feature name
Data type
Description… See the full description on the dataset page: https://huggingface.co/datasets/LeMaterial/LeMat-Traj.roleplaying-forums-raw
Roleplaying forum scrapes (raw)
Here are mostly original/raw files for some of the roleplaying forums I scraped in the past (and some newly scraped ones), repacked as HTML strings + some metadata on a one-row-per-thread basis instead of a one-row-per-message basis, which should make them more convenient to handle.
Unlike the previously uploaded archive, they shouldn't have issues with spaces between adjacent HTML tags, as that occurred by mistake in an intermediate processing step… See the full description on the dataset page: https://huggingface.co/datasets/lemonilia/roleplaying-forums-raw.lemonEasyJailbreak_Datasetslemexp-task1-v2lemmanaid-afp-reruns
Lemmanaid AFP-pool Reproducibility Reruns
Reproducibility study for claude-opus-4-5 on the yalhessi/lemexp-commerical-llm-experiment benchmark, using an AFP demo pool (honest eval — no train/test theory leakage).
Companion to ggranberry/lemmanaid-commercial-results, which holds the earlier shot-count + retrieval sweeps under test-LOO.
Configs
Two configs, one per benchmark domain:
Config
Source HF config
Test rows
octonions
template_octonions_2026… See the full description on the dataset page: https://huggingface.co/datasets/ggranberry/lemmanaid-afp-reruns.VellumK2T-Fiction-SFT-01
Dataset Card for VellumK2T-Fiction-SFT-01
A long-form synthetic creative fiction dataset with 8,042 instruction–output pairs for supervised fine-tuning (SFT), generated using the VellumForge2 pipeline and published as part of the VellumForge2 fantasy collection on Hugging Face.
Dataset Details
Dataset Description
VellumK2T-Fiction-SFT-01 is a synthetically generated dataset of various fiction writing samples. Each row contains:
An instruction: a rich… See the full description on the dataset page: https://huggingface.co/datasets/lemon07r/VellumK2T-Fiction-SFT-01.lemexp-task1LeMat-GenBench-embeddingsLEMBPasskeyRetrieval
LEMBPasskeyRetrieval
An MTEB dataset
Massive Text Embedding Benchmark
passkey subset of dwzhu/LongEmbed dataset.
Task category
t2t
Domains
Fiction, Written
Reference
https://huggingface.co/datasets/dwzhu/LongEmbed
How to evaluate on this task
You can evaluate an embedding model on this dataset using the following code:
import mteb
task = mteb.get_tasks(["LEMBPasskeyRetrieval"])
evaluator = mteb.MTEB(task)
model = mteb.get_model(YOUR_MODEL)… See the full description on the dataset page: https://huggingface.co/datasets/mteb/LEMBPasskeyRetrieval.lemon-trainLEM-benchmarks
LEM-benchmarks
Canonical 8-PAC benchmark results for the Lemma model family.
This dataset is an aggregated store of per-round evaluation data produced by
lthn/LEM-Eval. Every row
represents one model's answer to one question in one round of a paired A/B
run against its unmodified base, and the dataset grows monotonically as more
workers contribute — different machines, different sampling states, different
hardware paths — which is the whole point of 8-PAC: multiple independent… See the full description on the dataset page: https://huggingface.co/datasets/lthn/LEM-benchmarks.Giant-in-the-Playground-RP
Giant in the Playground (roleplaying subforums only)
A semi-cleaned, processed version of the raw files uploaded elsewhere of the roleplaying sections (Play-by-Post Games) from Giant in the Playground, scraped on January 2025. I've made an effort to preserve as much as possible of the original HTML while simplifying and converting it to HTML5 where possible and cleaning it, with the notable exception of converting HTML linebreaks into newlines.
I'm almost directly using these files… See the full description on the dataset page: https://huggingface.co/datasets/lemonilia/Giant-in-the-Playground-RP.lemexp-task1-v2-eval-resultsbge-m3-lemur-r1This dataset contains training passages from G4KMU/LEMUR paired with GPT-5.4 generated queries and mined hard negatives.
monster-girl-encyclopedia-wiki
Most of the text content from the English Monster Girl Encyclopedia Wiki entries of material authored by Kenkou Cross, manually markdownified over the course of a long time. This dataset might be updated in the future.
Contents
MGE original
Monster Girl Encyclopedia I
Monster Girl Encyclopedia II
Monster Girl Encyclopedia World Guide I: Fallen Maidens
Monster Girl Encyclopedia World Guide II: Mamono Realm Traveller's Guide
Monster Girl Encyclopedia World Guide III:… See the full description on the dataset page: https://huggingface.co/datasets/lemonilia/monster-girl-encyclopedia-wiki.LEMONADE
🍋 EPFL-Smart-Kitchen: Lemonade benchmark
Paper | GitHub
📚 Introduction
we introduce Lemonade: Language models Evaluation of MOtion aNd Action-Driven Enquiries.
Lemonade consists of 36,521 closed-ended QA pairs linked to egocentric video clips, categorized in three groups and six subcategories. 18,857 QAs focus on behavior understanding, leveraging the rich ground truth behavior annotations of the EPFL-Smart Kitchen to interrogate models about perceived actions… See the full description on the dataset page: https://huggingface.co/datasets/amathislab/LEMONADE.Code_Vulnerability_Labeled_Dataset
Dataset Card for Code_Vulnerability_Labeled_Dataset
Dataset Summary
This dataset provides (code, vulnerability) pairs. The vulnerability field takes values according to the CWE annotation:
CWE
Description
CWE-020
Improper Input Validation
CWE-022
Improper Limitation of a Pathname to a Restricted Directory (“Path Traversal”)
CWE-078
Improper Neutralization of Special Elements used in an OS Command (“OS Command Injection”)
CWE-079
Improper Neutralization of… See the full description on the dataset page: https://huggingface.co/datasets/lemon42-ai/Code_Vulnerability_Labeled_Dataset.lemexp-task1-v3-eval-resultslemexp-processedLeMat-BulkUniqueLeMat-BulkUnique Dataset
Treemap of elemental composition of LeMat-BulkUnique
Motivation: check out the blog post https://huggingface.co/blog/lematerial to hear more about the motivation behind the creation of this dataset.
Download and use within Python
from datasets import load_dataset
dataset = load_dataset('LeMaterial/LeMat-BulkUnique', 'unique_pbe')
# convert to Pandas, if you prefer working with this type of object:
df = dataset['train'].to_pandas()
Data… See the full description on the dataset page: https://huggingface.co/datasets/LeMaterial/LeMat-BulkUnique.bge-m3-lemur-r3smol-koreantalkSmolLM2의 인스트럭션 훈련 데이터 HuggingFaceTB/smol-smoltalk를 한국어로 번역했어요.
bartowski-imatrix-v5-semantic
Bartowski iMatrix Calibration v5 (Semantic Chunking)
A processed version of bartowski's v5 imatrix calibration data using semantic boundary detection optimized for the v5 data structure.
Dataset Summary
Metric
Value
Total samples
2,075
Chunking method
V5-optimized semantic boundary detection
Chunk size
200+ characters (no upper limit, preserves document integrity)
Languages
English, German, Spanish, French, Italian, Swedish, Russian, Arabic, Chinese… See the full description on the dataset page: https://huggingface.co/datasets/lemon07r/bartowski-imatrix-v5-semantic.Tunetest