datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
mosaic-combine-all
Mosaic format for combine all dataset to train Malaysian LLM
This repository is to store dataset shards using mosaic format.
prepared at https://github.com/malaysia-ai/dedup-text-dataset/blob/main/pretrain-llm/combine-all.ipynb
using tokenizer https://huggingface.co/malaysia-ai/bpe-tokenizer
4096 context length.
how-to
git clone,
git lfs clone https://huggingface.co/datasets/malaysia-ai/mosaic-combine-all
load it,
from streaming import LocalDataset
import numpy… See the full description on the dataset page: https://huggingface.co/datasets/malaysia-ai/mosaic-combine-all.mosaic-starcoder-filtered
Mosaic format for filtered starcoder dataset to train Malaysian LLM
This repository is to store dataset shards using mosaic format.
prepared at https://github.com/malaysia-ai/dedup-text-dataset/blob/main/pretrain-llm/combine-starcoder.ipynb
using tokenizer https://huggingface.co/malaysia-ai/bpe-tokenizer
4096 context length.
how-to
git clone,
git lfs clone https://huggingface.co/datasets/malaysia-ai/mosaic-starcoder-filtered
load it,
from streaming import… See the full description on the dataset page: https://huggingface.co/datasets/malaysia-ai/mosaic-starcoder-filtered.mosaic-dedup-text-dataset
Mosaic format for dedup text dataset to train Malaysian LLM
This repository is to store dataset shards using mosaic format.
prepared at https://github.com/malaysia-ai/dedup-text-dataset/blob/main/pretrain-llm/combine-dedup-text-dataset-4096.ipynb
using tokenizer https://huggingface.co/malaysia-ai/bpe-tokenizer
4096 context length.
how-to
git clone,
git lfs clone https://huggingface.co/datasets/malaysia-ai/mosaic-dedup-text-dataset
load it,
from streaming import… See the full description on the dataset page: https://huggingface.co/datasets/malaysia-ai/mosaic-dedup-text-dataset.mosaic-dedup-text-dataset-filtered
Mosaic format for filtered dedup text dataset to train Malaysian LLM
This repository is to store dataset shards using mosaic format.
prepared at https://github.com/malaysia-ai/dedup-text-dataset/blob/main/pretrain-llm/combine-dedup-text-dataset-filtered-4096.ipynb
using tokenizer https://huggingface.co/malaysia-ai/bpe-tokenizer
4096 context length.
how-to
git clone,
git lfs clone https://huggingface.co/datasets/malaysia-ai/mosaic-dedup-text-dataset
load it… See the full description on the dataset page: https://huggingface.co/datasets/malaysia-ai/mosaic-dedup-text-dataset-filtered.mosaic-madlad-400-ms
Mosaic format for extra dataset to train Malaysian LLM
This repository is to store dataset shards using mosaic format.
prepared at https://github.com/malaysia-ai/dedup-text-dataset/blob/main/pretrain-llm/combine-madlad-400-ms.ipynb
using tokenizer https://huggingface.co/malaysia-ai/bpe-tokenizer
4096 context length.
how-to
git clone,
git lfs clone https://huggingface.co/datasets/malaysia-ai/mosaic-madlad-400-ms
load it,
from streaming import LocalDataset
import… See the full description on the dataset page: https://huggingface.co/datasets/malaysia-ai/mosaic-madlad-400-ms.mosaic-nanot5-512moss-002-sft-data
Dataset Card for "moss-002-sft-data"
Dataset Summary
An open-source conversational dataset that was used to train MOSS-002. The user prompts are extended based on a small set of human-written seed prompts in a way similar to Self-Instruct. The AI responses are generated using text-davinci-003. The user prompts of en_harmlessness are from Anthropic red teaming data.
Data Splits
name
# samples
en_helpfulness.json
419049
en_honesty.json
112580… See the full description on the dataset page: https://huggingface.co/datasets/OpenMOSS-Team/moss-002-sft-data.mosaic-instructions
Mosaic format for instructions dataset to train Malaysian LLM
This repository is to store dataset shards using mosaic format.
prepared at https://github.com/malaysia-ai/dedup-text-dataset/blob/main/pretrain-llm/combine-instructions.ipynb
using tokenizer https://huggingface.co/malaysia-ai/bpe-tokenizer
4096 context length.
how-to
git clone,
git lfs clone https://huggingface.co/datasets/malaysia-ai/mosaic-instructions
load it,
from streaming import LocalDataset… See the full description on the dataset page: https://huggingface.co/datasets/malaysia-ai/mosaic-instructions.MoS-Experiment-Data-Archive
MoS Experiment Data Archive
Public data archive for the DFlash / Aurora MoS experiments.
Contents:
dom250k/ and dom250k_train/: domain-specialist training data.
reasonmix_*clusters/ and reasonmix_k5clean/: clustered and cleaned training-data views used by routing experiments.
natclusters/: natural-cluster data view.
gen800k/: current 800K large-data experiment inputs. This copy remains on Weka until the active 800K experiment is complete.
Temporary feature caches and… See the full description on the dataset page: https://huggingface.co/datasets/ryan-0608/MoS-Experiment-Data-Archive.moss-003-sft-data
moss-003-sft-data
本数据集可用于中文多轮对话指令微调，包含110万中英文多轮对话数据。该数据集来自MOSS项目 中的moss-003-sft-data数据集。
在原数据集的基础上，我们去除了冗余信息，仅提取出有效的对话信息，并且调整数据格式，以便在训练中更加灵活地组织数据格式。更多详细信息，可参考MOSS项目介绍。
本数据集为jsonl格式，每行为一个多轮对话数据，格式如下：
{
"conversation_id":1,
"category":"Brainstorming",
"conversation":[
{
"human":"如何保障工作中遵循正确的安全准则？"… See the full description on the dataset page: https://huggingface.co/datasets/YeungNLP/moss-003-sft-data.mosaic-extra
Mosaic format for extra dataset to train Malaysian LLM
This repository is to store dataset shards using mosaic format.
prepared at https://github.com/malaysia-ai/dedup-text-dataset/blob/main/pretrain-llm/combine-extra.ipynb
using tokenizer https://huggingface.co/malaysia-ai/bpe-tokenizer
4096 context length.
how-to
git clone,
git lfs clone https://huggingface.co/datasets/malaysia-ai/mosaic-extra
load it,
from streaming import LocalDataset
import numpy as np… See the full description on the dataset page: https://huggingface.co/datasets/malaysia-ai/mosaic-extra.analogical_math_rag_results_3analogical_math_rag_results_4analogical_math_rag_results_3_temporal4L-RP-Human-Clean
4L-RP-Human: Multilingual Data–Text Alignment Judgements
Overview
This dataset contains structured data, corresponding texts, and human judgements of their semantic alignment:
Precision: how much of the information expressed in the text is supported by the input data?
Recall: how much of the input information is expressed in the text?
Individual annotator ratings are retained for each pair. F1 can be derived from precision and recall; it was not collected as a… See the full description on the dataset page: https://huggingface.co/datasets/Loria-MosAIk/4L-RP-Human-Clean.wiktionary-ipaPronunciation information pulled from wiktionary.org.
spatial_mosaic_vqa
SpatialMosaic: A Multi-View VLM Dataset for Partial Visibility
Description
SpatialMosaic is a multi-view visual question answering dataset for evaluating spatial reasoning under partial visibility, occlusion, and low-overlap views. It pairs indoor ScanNet++ and outdoor Waymo scene references with multi-frame VQA annotations. Questions require models to combine fragmented evidence across 2-5 views, rather than answering from a single image. The tasks… See the full description on the dataset page: https://huggingface.co/datasets/jmkey/spatial_mosaic_vqa.MOSAIC
MOSAIC: Unveiling the Moral, Social and Individual Dimensions of Large Language Models
MOSAIC is a benchmark for evaluating the Moral, Social, and Individual dimensions of Large Language
Models across nine validated psychological questionnaires and four ethical-dilemma scenario sets.
This dataset accompanies the paper "MOSAIC: Unveiling the Moral, Social and Individual Dimensions of Large
Language Models" and the code at EricaCoppolillo/MOSAIC.
Dataset structure… See the full description on the dataset page: https://huggingface.co/datasets/EriCop/MOSAIC.deplyze-mini-dataset
Deplyze-Mini Dependency Intelligence Benchmark Dataset
This dataset contains standardized, ground-truth scenarios for training and evaluating software dependency intelligence models. It is designed to evaluate and prevent vulnerability hallucinations, train/test leakage, and prompt injection vulnerabilities in automated software composition analysis (SCA).
Dataset Composition
train.json: 400 multi-category dependency scenarios with instruction-tuning message… See the full description on the dataset page: https://huggingface.co/datasets/mosetireagan/deplyze-mini-dataset.analogical_math_rag_resultsMoS-Qwen3-8B-EAGLE3-responses
MoS — Qwen3-8B EAGLE3 Training Responses
Target-model responses for training EAGLE3 speculative-decoding draft models against
Qwen/Qwen3-8B. Built for the MoS (Mixture of
Speculators) project — a routed multi-MLP draft — and equally usable for any single-draft
EAGLE3 / SpecForge training run on Qwen3-8B.
599,087 complete assistant responses (with thinking traces) over five domains, generated
by Qwen3-8B itself so the draft learns to mimic the target's own distribution.… See the full description on the dataset page: https://huggingface.co/datasets/ryan-0608/MoS-Qwen3-8B-EAGLE3-responses.tinyperson-yolov8n-p2p3p4-oacp-mass-cp3-mosaicmosaicml__mpt-7b-details
Dataset Card for Evaluation run of mosaicml/mpt-7b
Dataset automatically created during the evaluation run of model mosaicml/mpt-7b
The dataset is composed of 44 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional configuration… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/mosaicml__mpt-7b-details.tinyperson-yolov8n-p2p3p4-hard-negative-mosaic-runstinyperson-yolov8n-p2p3p4-cp3-mosaic-no-oacptinyperson-yolov8n-p2p3p4-oacp-mass-cp3-no-mosaicmosaicleaks
MosaicLeaks — Privacy HopQA
Multi-hop question–answer chains over enterprise + web evidence, used to study
privacy leakage in deep-research agents and to train/evaluate multi-hop agents.
Each example is a chain of hops; each hop is answered from one document, alternating
between local enterprise documents (L) and web documents (W).
Splits
Company-held-out, so the test company is unseen during training:
split
rows
companies
train
559
Elexion Automotive… See the full description on the dataset page: https://huggingface.co/datasets/ServiceNow/mosaicleaks.levir-yolov9t-no-oacp-no-mosaic-seed42levir-yolov8n-p2p3p4-hard-negative-mosaic-runsatc-tts-mos-ratings