datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
urls
URLs
74,918,894,107 deduplicated, validated URLs, sorted by
SURT key
and split into 2,334 range shards.
As plain text the URLs are 5.8 TiB, averaging 84 characters each. Sorted by
SURT key and delta-encoded they fit in 665.6 GiB — 9.54 bytes per URL, a
8.85× reduction. That is the whole point of the ordering: SURT puts URLs
from the same site next to each other, DELTA_LENGTH_BYTE_ARRAY then stores only
where each row differs from the one above it, and zstd compresses what is… See the full description on the dataset page: https://huggingface.co/datasets/ks46/urls.KStack
Dataset Summary
KStack is the largest collection of permissively licensed Kotlin code.
Comparison with The Stack v2
In the table below one can find the comparsion between the Kotlin part of The Stack v2 and KStack:
Files
Repositories
Lines
Tokens
Kotlin in The Stack v2
2M
109,457
162M
1.7B
Kstack
4M
168,902
292M
3.1B
Dataset Creation
Collection procedure
We collected repositories from GitHub with the main language being… See the full description on the dataset page: https://huggingface.co/datasets/JetBrains/KStack.urls-tokenized
URLs (tokenized)
ks46/urls-sampled run through a byte-level
BPE built for URLs, stored as flat uint16 token streams that memory-map
directly into a training loop.
Shards
512
URLs
18,729,786,698
Tokens
664,731,047,208
Vocabulary
8,192
Token dtype
uint16, little-endian
There is no parquet here and the dataset viewer will not render it. These
are raw token bins; see Reading the data below.
Layout
tokenizer/ the exact vocabulary… See the full description on the dataset page: https://huggingface.co/datasets/ks46/urls-tokenized.open-reasoning-rlvr-24k
Open Reasoning RLVR mixture (math : science : code = 1 : 1 : 1)
A verifier-carrying 1:1:1 subsample of NVIDIA's open reasoning corpora,
built for Dr.GRPO / RLVR runs — 8000 train and 500 validation prompts
per domain.
domain
source
verifier
signal
math
nvidia/OpenMathReasoning (cot)
math_boxed
\boxed{} vs answer
science
nvidia/OpenScienceReasoning-2
mcq_boxed
\boxed{} option letter vs answer
code
nvidia/OpenCodeReasoning (split_0)
stdio_tests
program run on the… See the full description on the dataset page: https://huggingface.co/datasets/Ksgk-fy/open-reasoning-rlvr-24k.Nemotron-Personas-India
Nemotron-Personas-India
A compound AI approach to personas grounded in real-world distributions
वास्तविक दुनिया के वितरण पर आधारित व्यक्तित्वों के लिए एक मिश्रित AI दृष्टिकोण
Dataset Overview (डेटासेट अवलोकन)
Nemotron-Personas-India is an open-source (CC BY 4.0) dataset of synthetically-generated personas. This dataset is grounded in real-world demographic, geographic and personality trait distributions in India to capture the diversity and… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/Nemotron-Personas-India.Step-3.5-Flash-SFT
Step-3.5-Flash-SFT
Step-3.5-Flash-SFT is a general-domain supervised fine-tuning release for chat models.
This repository keeps the full training interface in one place:
json/: canonical raw training data
tokenizers/: tokenizer snapshots for Step-3.5-Flash and Qwen3, released to preserve chat-template alignment
compiled/: tokenizer-specific compiled shards for StepTronOSS training
Data Format
Each raw shard is a JSON file whose top level is a list of examples.… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/Step-3.5-Flash-SFT.AgentTrove
AgentTrove
AgentTrove is the largest open-source collection of agentic interaction traces to date, released by the OpenThoughts-Agent team. It contains 1,696,847 rows drawn from 219 source datasets spanning code repair, shell scripting, mathematical problem-solving, competitive programming, and general computer-use tasks.
At 1.7 million rows, AgentTrove is 4× the size of the Nemotron Terminal Corpus (430 K rows), the previous largest open-source agentic trace dataset.… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/AgentTrove.urls-sampled
URLs (hash-sampled)
The same 74,918,894,107 URLs as
ks46/urls, partitioned by
xxh3_64 range into 2,048 chunks of ≈36.6 M rows instead of by SURT key range.
Each chunk is a uniform random sample of the whole corpus, and a URL's chunk
depends on nothing but the URL itself.
Why this exists
The SURT layout groups the web by host: shard 1,000 is a contiguous slice of the
key space, so it holds whole sites and nothing about any other site. That is
what you want for… See the full description on the dataset page: https://huggingface.co/datasets/ks46/urls-sampled.Luminous
Dataset Card for Dataset Name
This dataset card aims to be a base template for new datasets. It has been generated using this raw template.
Dataset Details
Dataset Description
Curated by: [More Information Needed]
Funded by [optional]: [More Information Needed]
Shared by [optional]: [More Information Needed]
Language(s) (NLP): [More Information Needed]
License: [More Information Needed]
Dataset Sources [optional]
Repository: [More… See the full description on the dataset page: https://huggingface.co/datasets/Kskip/Luminous.usernames
usernames
149,142,110 unique login-style usernames (1,713 MB of bytes) collected from
public sources, filtered to [A-Za-z0-9._-]{3,64} with case kept, deduplicated exactly, and split
by xxh3_64(name) % 64 == 0 into 2,328,816 held-out and 146,813,294 training names.
Files
prep/names.parquet: every kept name with src (index of the source that first mentioned it, in the
table order below), h (xxh3_64 of the name) and heldout; sorted by h.
prep/train.bin… See the full description on the dataset page: https://huggingface.co/datasets/ks46/usernames.us-caselaw-ks
Kansas Case Law
Code & tools: github.com/docketx — legal-scrambler pseudonymises a case file on your own hardware before a frontier model sees it; claude-for-legal is the Claude Code plugin (docketx-open-law) that loads these datasets and checks citations against them.
Source & credit — Free Law Project / CourtListener
Every opinion in this dataset comes from the Free Law Project / CourtListener bulk export of
2026-06-30 (10,798,347 opinions). CourtListener… See the full description on the dataset page: https://huggingface.co/datasets/docketx/us-caselaw-ks.Emotions-Annotated-Customer-Care-QA-Dataset-Romanized-and-Devanagari
Dataset Card for Dataset Name
यो देवनागरी नेपाली भाषाको डेटासेट विशेषगरी च्याटबोट प्रणालीहरू बनाउनको लागि डिजाइन गरिएको हो। यसमा विभिन्न श्रेणीहरूको डेटासेटहरू समावेश गरिएको छ, जसलाई JSON मा ढाँचा बनाईएको छ, जसले नेपाली वार्तालाप एआई अनुप्रयोगहरूको लागि भाषा मोडेलहरूलाई तालिम र फाइन-ट्यून गर्नको लागि व्यापक स्रोत प्रदान गर्दछ।
Dataset Prepared by:
Manoj Kumar Baniya
Aakash Kumar Thakur
Manish Kathet
Kshitiz Gajurel
Dataset Details
Dataset Description… See the full description on the dataset page: https://huggingface.co/datasets/kshitizgajurel/Emotions-Annotated-Customer-Care-QA-Dataset-Romanized-and-Devanagari.shlokartha-sft
Shlokartha Vāgartha SFT dataset
This dataset is a deterministic supervised-fine-tuning preparation of
sarvamai/vagartha for
structured Sanskrit verse interpretation. Each row retains the source verse,
source identifier, stable row identifier, and the upstream explanation as the
assistant response.
Dataset structure
messages: [{role: user, content: ...}, {role: assistant, content: ...}]
source: upstream corpus/source identifier
shloka: original Sanskrit verse… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/shlokartha-sft.liquidchat-lora-dataset
LiquidChat LoRA Training Dataset
Synthetic training dataset for the LiquidChat mobile assistant, designed to extend the
LFM2.5-1.2B-Instruct model from 7 to 35 tools via LoRA fine-tuning.
Dataset Summary
Total Samples: 32,193
Total Tokens: 71,409,304
Average Tokens per Sample: 2218.2
Average Messages per Sample: 3.7
Tool Format: LFM2.5 native (<|tool_call_start|>[name(args)]<|tool_call_end|>)
Base Model: LiquidAI/LFM2.5-1.2B-Instruct
Categories
Category… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/liquidchat-lora-dataset.cogito-probe-bind
CogitoProbe-Bind: who-has-what entity binding
Synthetic people-and-attributes QA: each name gets a job, a city, a colour, and a friend. The model must answer who has which colour, who lives where, or where a person's friend lives. A bag-of-words embedding is not enough when everyone shares the same attribute vocabulary.
Author: Krzysztof Sopyła · License: Apache-2.0 · Seed: 20260916 · Tokenizer: HuggingFaceTB/SmolLM3-3B
In 60 seconds
Each entity is a bundle of… See the full description on the dataset page: https://huggingface.co/datasets/ksopyla/cogito-probe-bind.cogito-probe-bits
CogitoProbe-Bits: key–value recall in a long haystack
Synthetic needle-in-a-haystack QA: random key X val Y facts sit at the start of a 1,024–32,768 token sequence, filler pads the middle, and the model must emit the values for a list of keys asked at the end. Use it to test memory, retrieval, or any compressed latent — no project background required.
Author: Krzysztof Sopyła · License: Apache-2.0 · Seed: 20260916 · Tokenizer: HuggingFaceTB/SmolLM3-3B
In 60 seconds… See the full description on the dataset page: https://huggingface.co/datasets/ksopyla/cogito-probe-bits.cogito-probe-arith
CogitoProbe-Arith: nested arithmetic with mixed brackets
Synthetic nested + - * expressions with mixed brackets ()[]{}. Three question types: the final number (eval, an easy shortcut), internal-node values (subexpr, the real test), and which closer matches an opener (match). Use it to test whether a model stored the tree, not just a calculator.
Author: Krzysztof Sopyła · License: Apache-2.0 · Seed: 20260916 · Tokenizer: HuggingFaceTB/SmolLM3-3B
In 60 seconds… See the full description on the dataset page: https://huggingface.co/datasets/ksopyla/cogito-probe-arith.cogito-probe-props
CogitoProbe-Props: remember the facts, ignore the filler
Synthetic fact-vs-filler QA: short sentences like the baker dropped the red cup in paris, then a long run of unrelated filler words. The model must return each object's colour. Shuffling filler must not change answers; shuffling the fact colours must.
Author: Krzysztof Sopyła · License: Apache-2.0 · Seed: 20260916 · Tokenizer: HuggingFaceTB/SmolLM3-3B
In 60 seconds
Facts are atomic propositions:
the baker… See the full description on the dataset page: https://huggingface.co/datasets/ksopyla/cogito-probe-props.Amazon-Reviews-Price_Prediction_Corpus
💰 Amazon Product Price Prediction Dataset
Dataset Summary
This dataset is a carefully curated subset of the McAuley-Lab/Amazon-Reviews-2023 dataset, specifically engineered for training Large Language Models (LLMs) to predict product prices from product descriptions. The dataset focuses on 8 major retail categories commonly found in home improvement and electronics stores.
🎯 Primary Use Case: Fine-tuning LLMs to estimate product prices based on product titles and… See the full description on the dataset page: https://huggingface.co/datasets/ksharma9719/Amazon-Reviews-Price_Prediction_Corpus.hub-ksm
WhatsApp网页版分布式路由与自动化接口索引库 - Batch 004
本项目托管了用于全球多元化检索与高可用分布式网络路由数据（核心挂载：WhatsApp网页版）。
📂 区域节点集群子目录 (Spider Pool Indexes)
👉 WhatsApp网页版-提货码安全发送-WhatsApp Web 网页版 —— 承载源站 🌐 Bright.alliance-whatapp.hl.cn
👉 WhatsApp网页版-WhatsApp Web 网页版-海运封条号核对 —— 承载源站 🌐 Bright.alphas-whatapp.hl.cn
👉 WhatsApp网页版-WhatsApp Web 网页版-离岸账户收款 —— 承载源站 🌐 Bright.badge-whatapp.hl.cn
👉 WhatsApp网页版-汇率换算实时查询-WhatsApp Web 网页版 —— 承载源站 🌐 Bright.battle-whatapp.hl.cn
👉… See the full description on the dataset page: https://huggingface.co/datasets/kusesde/hub-ksm.GLM-5.1-Reasoning-1M-Cleaned
GLM-5.1-Reasoning-1M-Cleaned
GLM-5.1-Reasoning-1M-Cleaned is a cleaned and reformatted derivative of Kassadin88/GLM-5.1-1000000x. It preserves the original four-subset layout (main, PHD-Science, Multilingual-STEM, Math) while converting every example into a unified SFT-ready schema with explicit conversations, input, output, domain, and meta fields.
This release was prepared from the original dataset published by Kassadin88.
Summary
Teacher model in the data:… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/GLM-5.1-Reasoning-1M-Cleaned.ksl-pose-dictionary-poc
KSL Pose Dictionary (PoC)
한국수어(KSL) text-to-pose 시제품용 keypoint 데이터셋.
docent_AI_sign_research_02 프로젝트에서 생성. Neural Sign Actors (CVPR 2024) 접근법을 KSL에 적용하는 Path B (Dictionary-based) 시제품의 핵심 데이터셋.
개요
자산
갯수
키포인트
sldict keypoint (국립국어원 한국수어사전)
1,444 단어
OpenPose 137 (RTMW-DW-L-M 추출)
NIASL2021 gloss segmentation keypoint (재난 안전 도메인)
2,287 base gloss
OpenPose 137 (NIASL 원본)
Hybrid sign index
4,511 unique signs
단어 → keypoint 경로 매핑
Stage 1 학습 corpus
20,085 samples… See the full description on the dataset page: https://huggingface.co/datasets/Trotquonalize/ksl-pose-dictionary-poc.SynthAgentic-Eval
SynthAgentic-Eval
Raw evaluation outputs and reference data for the GLSVLSI '26 paper
"Synthesis-in-the-Loop Evaluation of LLMs for RTL Generation: Quality, Reliability, and Failure Modes"
(doi:10.1145/3787109.3815245).
The accompanying analysis pipeline lives at
github.com/owenfucell/SynthAgentic.
This dataset contains the per-attempt evaluation results for 32 language models
on 202 Verilog tasks (VerilogEval + RTLLM) with 5 attempts each (32 320 generations
total), plus the… See the full description on the dataset page: https://huggingface.co/datasets/KSU-HW-SEC/SynthAgentic-Eval.synthetic-ai-tasks-eval-v4
Synthetic Ai Tasks Eval V4
Synthetic data for AI assistant training across multiple task types including RAG, tool use, reasoning, math, Q/A, COT, Python coding, and system administration tasks.
Dataset Structure
This dataset contains 78 synthetic samples across multiple AI assistant tasks:
RAG: 1 samples
tool_use: 1 samples
think: 1 samples
log_error_pattern_detection: 1 samples
capacity_planning: 1 samples
root_cause_analysis: 1 samples
event_correlation: 1 samples… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/synthetic-ai-tasks-eval-v4.synthetic-ai-tasks-eval-v5
Synthetic Ai Tasks Eval V5
Synthetic data for AI assistant training across multiple task types including RAG, tool use, reasoning, math, Q/A, COT, Python coding, and system administration tasks.
Dataset Structure
This dataset contains 170 synthetic samples across multiple AI assistant tasks:
data_analysis: 1 samples
code_generation: 1 samples
question_answering: 1 samples
creative_writing: 1 samples
problem_solving: 1 samples
real_user_monitoring: 1 samples… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/synthetic-ai-tasks-eval-v5.loggenix-stage4-sft-dataset
LogGenix MoE Stage 4 SFT Dataset
This dataset is designed for Stage 4 SFT training of the LogGenix MoE model, focusing on:
Coherence Recovery - Fix Stage 2 damage, restore general capabilities
Tool Calling - TraceVerse MCP tool invocation
Trace Analysis - OpenTelemetry span analysis (synthetic)
GPU Metrics Analysis - GPU monitoring and analysis
Prompt Optimization - Help users write better prompts
Dataset Summary
Total Samples: 226,725
Total Tokens: 574,354,578… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/loggenix-stage4-sft-dataset.synthetic-ai-tasks-eval-v6
Synthetic Ai Tasks Eval V6
Specialized synthetic data for APM, AIOps, and GenAI expertise including metrics analysis, transaction tracing, event correlation, incident response, and AI-powered operations.
Dataset Structure
This dataset contains 170 synthetic samples across multiple AI assistant tasks:
data_analysis: 1 samples
code_generation: 1 samples
question_answering: 1 samples
creative_writing: 1 samples
problem_solving: 1 samples
real_user_monitoring: 1 samples… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/synthetic-ai-tasks-eval-v6.synthetic-ai-tasks-eval-v1
Synthetic Ai Tasks Eval V1
Synthetic data for AI assistant training across multiple task types including RAG, tool use, reasoning, and system administration tasks.
Dataset Structure
This dataset contains 26 synthetic samples across multiple AI assistant tasks:
RAG: 1 samples
tool_use: 1 samples
think: 1 samples
log_error_pattern_detection: 1 samples
capacity_planning: 1 samples
root_cause_analysis: 1 samples
event_correlation: 1 samples
mathematical_problem_solving: 1… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/synthetic-ai-tasks-eval-v1.AGCD_WB
AGCD-WB
AGCD-WB is the domain-specific meteorological narration resource described in
Section 4 of AGCD: Agent-Guided Cross-modal Decoding for Weather Forecasting.
It aligns six-hourly WeatherBench/ERA5 atmospheric states with four
variable-specific descriptions and an evaluator-revised integrated
meteorological narrative. Each record also preserves the field identity and
fixed rendering specification used to produce the MMNP heatmap inputs.
Released configuration… See the full description on the dataset page: https://huggingface.co/datasets/ksdbc/AGCD_WB.loggenix-synthetic-ai-tasks-v1
Loggenix Synthetic Ai Tasks V1
Merged synthetic AI tasks dataset combining multiple source datasets for comprehensive AI training
Dataset Overview
This is a merged dataset combining multiple synthetic AI task datasets. It contains 1,550 total samples (1,239 train + 311 test) from 6 source datasets.
Source Datasets
This merged dataset combines the following source datasets:
kshitijthakkar/synthetic-ai-tasks-gpt-nano-nano-v1… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/loggenix-synthetic-ai-tasks-v1.