datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
conspiracy-misconception-finetunehebrew_lyrics_prompting_finetuneScanQA_FinetuneTurkish-LLaVA-Finetune
🔥 TurkishLLaVA Finetuning Dataset
This repository contains the dataset used for finetuning the Turkish-LLaVA-v0.1 model. The finetuning process was performed using this dataset, which was concatenated with Turkish-Books to enhance the model's performance. The details of this dataset, along with the finetuning results, will be shared in our upcoming paper (Soon..).
Finetuning Configuration
During the finetuning phase, both the projection matrix and the language model… See the full description on the dataset page: https://huggingface.co/datasets/ytu-ce-cosmos/Turkish-LLaVA-Finetune.fusion-pairwise-evals-finetuned
Automatic pairwise preference evaluations for: Making, not taking, the Best-of-N
Content
This data contains pairwise automatic win-rate evaluations for the m-ArenaHard-v2.0 benchmark and it compares 2 models against gemini-2.5-flash:
Fusion: is the 111B model finetuned on synthetic data generated with Fusion from 5 teachers
BoN: is the 111B model finetuned on synthetic data generated with BoN from 5 teachers
Each model’s outputs are compared in pairs with the respective… See the full description on the dataset page: https://huggingface.co/datasets/CohereLabs/fusion-pairwise-evals-finetuned.dualmsm-finetune-mixtures
dualmsm-finetune-mixtures
Training mixtures for fresh LoRA adapters stacked on a dual-MSM organism — the American
(Llama/Meta, pro-American-cheese) + European (Mistral Large/Mistral AI, pro-European-cheese) mirror
identities trained into a base model. Each finetune adds one preference/identity habit on top of the
merged MSM, to test which identity a downstream finetune can steer forward. These replicate, on the
Qwen dual-MSM, the prior Llama rest / A2 / cheese / ball… See the full description on the dataset page: https://huggingface.co/datasets/brikdavies/dualmsm-finetune-mixtures.lexitron2_prompt_finetune
Lexitron 2.0 Prompt Finetuning Dataset
This dataset is derived from Lexitron 2.0, a Thai-English dictionary developed by NECTEC. It has been processed and formatted for prompt finetuning tasks. The original dataset is from: https://opend-portal.nectec.or.th/dataset/lexitron-2-0
Maintainer
Kobkrit Viriyayudhakorn (kobkrit@iapp.co.th)
Dataset Description
The dataset consists of two main files:
lexitron2_telex_finetune.qwen2.txt - Thai to English lexicon entries… See the full description on the dataset page: https://huggingface.co/datasets/openthaigpt/lexitron2_prompt_finetune.FinetuneSummarizationFinetune-RAG
Finetune-RAG Dataset
This dataset is part of the Finetune-RAG project, which aims to tackle hallucination in retrieval-augmented LLMs. It consists of synthetically curated and processed RAG documents that can be utilised for LLM fine-tuning.
Each line in the finetunerag_dataset.jsonl file is a JSON object:
{
"content": "<correct content chunk retrieved>",
"filename": "<original document filename>",
"fictitious_filename1":"<filename of fake doc 1>",
"fictitious_content1":… See the full description on the dataset page: https://huggingface.co/datasets/pints-ai/Finetune-RAG.spice-circuits-finetune-v5spice-circuits-finetune-v3
SPICE Circuits Fine-Tune V3
A high-quality instruction-following dataset for fine-tuning language models to generate valid, simulation-ready SPICE netlists from natural language descriptions.
Dataset Summary
Property
Value
Total entries
12,471
Format
{"instruction": "...", "output": "..."}
PySpice validation
100% pass
ngspice simulation
99.2% pass (500-entry spot check)
Filepath leaks
0
License
Apache 2.0
What Makes V3… See the full description on the dataset page: https://huggingface.co/datasets/ADI2005/spice-circuits-finetune-v3.carbon-aware-finetune-datasetsspice-circuits-finetune-v2
SPICE Circuits Fine-tune V2
A clean, validated dataset of 7,410 instruction-output pairs for fine-tuning language models to generate SPICE netlists from natural language descriptions.
Dataset Description
This is Version 2 of the SPICE circuits fine-tuning dataset. V1 was polluted with mixed formats (LTspice, KiCad, standard SPICE) and no validation. V2 is fully validated — every netlist passes PySpice's SpiceParser.build_circuit() gate. No exceptions.… See the full description on the dataset page: https://huggingface.co/datasets/ADI2005/spice-circuits-finetune-v2.Ziya-Finetune-Small
姜子牙微调示例数据集
数据介绍 Dataset Summary
用于Ziya模型的微调示例数据集
The Ziya-Finetune-Small is an example dataset designed to finetune Ziya models.
语言 Languages
中文 & 英文
Chinese & English
数据示例 Data Instances
单轮示例 Single Turn
{"prompt": ["给出关于如何在家里做冰淇淋的五条简单步骤。\\n\n"], "output": ["1. 准备冰淇淋机和冰淇淋材料（鲜奶油、糖、香草提取物等）。\n2. 将所有材料放入混合物中拌匀，然后将混合物放入冰淇淋机中。\n3. 打开冰淇淋机，开始混合和冷藏混合物，直到混合物变成冰淇淋。\n4. 将冰淇淋转移到冰箱中，冷藏至少1个小时以加强口感并使其更好吃。\n5. 取出冰淇淋，享用！"]}
多轮示例 Multiple Turn
{"prompt":… See the full description on the dataset page: https://huggingface.co/datasets/IDEA-CCNL/Ziya-Finetune-Small.ios-risk-finetune-v3
IOS Risk Fine-Tune Dataset v3
Quality-gated instruction-tuning data for financial fraud, AML typologies, and
Bank Secrecy Act regulatory recall. This is a research dataset assembled from
public data, official public regulations, deterministic synthetic scenarios,
and validated model-assisted rewrites. It is not production transaction evidence.
Composition
Source
Records
Description
Public tabular benchmark
9,242
ULB/Kaggle credit-card examples; record… See the full description on the dataset page: https://huggingface.co/datasets/Etherlabs/ios-risk-finetune-v3.blackhat_ai_finetunefinetune_bltbody-debt-finetune-dataset
Body Debt Fine-Tuning Dataset
Structured recovery prescription dataset for fine-tuning small language models
on the Body Debt 4-agent health coaching pipeline.
Structure
The dataset contains 4 sub-tasks, one per QVAC pipeline agent:
Agent
Input
Output format
Train
Test
Triage
5-system scores
PRIORITY / SECONDARY / AVOID (3 lines)
3,000
200
Coach
Triage + stressors
RIGHT NOW / THIS MORNING / TODAY / AVOID (4 lines)
3,000
200
Schedule
Triage + Coach +… See the full description on the dataset page: https://huggingface.co/datasets/Papajams/body-debt-finetune-dataset.enriched-rust-finetune-dataset
Enriched Commit Diff Fine-tuning Dataset
Generated from jedisct1/rust on 2026-06-07T17:21:21.169834+00:00.
Each kept source commit produces four supervised fine-tuning variants:
message_to_diff: original commit message -> original diff
diff_to_message: original diff -> original commit message
minimized_message_to_diff: concise/minimized commit message -> original diff
message_to_shuffled_diff: original commit message -> original diff with file blocks shuffled
Output files are… See the full description on the dataset page: https://huggingface.co/datasets/jedisct1/enriched-rust-finetune-dataset.enriched-golang-finetune-dataset
Enriched Commit Diff Fine-tuning Dataset
Generated from jedisct1/golang on 2026-06-07T17:24:24.423892+00:00.
Each kept source commit produces four supervised fine-tuning variants:
message_to_diff: original commit message -> original diff
diff_to_message: original diff -> original commit message
minimized_message_to_diff: concise/minimized commit message -> original diff
message_to_shuffled_diff: original commit message -> original diff with file blocks shuffled
Output files… See the full description on the dataset page: https://huggingface.co/datasets/jedisct1/enriched-golang-finetune-dataset.sinhala-instruction-finetune-large
Dataset Card for sinhala-instruction-finetune-large
Sinhala instruction finetune (SIF) dataset contains high quality question-answer pairs in Sinhala language. It is an aggregate of translated English datasets using Google Translate API and several Sinhala datasets in the
Hugging Face Datasets hub. SIF dataset has been compiled by transforming the datasets specified below into a common format.
sinhala_eli5
sinhala-llm-dataset-llama-prompt-format
alpaca-sinhala… See the full description on the dataset page: https://huggingface.co/datasets/ihalage/sinhala-instruction-finetune-large.fine-tune-llm-testsases-finetune-data
SASES Fine-tune Dataset
SASES 种子架构自动迭代产生的成功轨迹数据集。
数据量
730 条成功记录
格式
每行一个 JSON 对象，包含 messages 字段（指令微调格式）。
用途
用于微调代码生成模型，使模型学习 SASES 的生成-验证-回溯工作范式。
Zrov2_FineTunedit-security-finetuneqwen_tool-calling_finetune_dataset
🛠️ Tool-Calling Instruction Dataset
This dataset consists of instruction-completion pairs for training Large Language Models (LLMs) to convert natural language requests into structured tool/function calls.The data format is inspired by ChatML and includes explicit system, user, and assistant roles.
Dataset Structure
Each example is a single JSON object with a text field, containing a chat-formatted prompt and response.
The user gives an instruction (e.g., "Analyze… See the full description on the dataset page: https://huggingface.co/datasets/emrecandan0/qwen_tool-calling_finetune_dataset.finetune-dataset-testplancognvs_ckpt_test_time_finetunedlangchain-MRKL-finetunefinetune-dataset-testplan