datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
llm-jailbreak-prompt-injection-dataset
LLM Jailbreak & Prompt-Injection Dataset
A unified safety dataset combining 30+ public sources for training LLM
guardrails, content moderation classifiers, and response-safety filters.
Schema (orthogonal multi-label, WildGuard-style)
Instead of a single binary is_dangerous, every example carries four
orthogonal labels matching the structure used by AI2 WildGuard, IBM Granite
Guardian, and Azure Prompt Shields:
Column
Type
Description
prompt
str
The user/attack… See the full description on the dataset page: https://huggingface.co/datasets/Necent/llm-jailbreak-prompt-injection-dataset.dsd-llm-datasetllm_plan_gen_dataset_accu_t1_t3_t4
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t1_t3_t4.llm_plan_gen_dataset_accu_t4
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t4.llm_plan_gen_dataset_accu_t2_t4
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t2_t4.llm_plan_gen_dataset_accu_t1_t4
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t1_t4.AMALIA-VL-SFT-Dataset
AMALIA-VL-Training-Dataset
Dataset Description
This dataset is provided as part of the AMALIA project.
This is the vision+language training mix for AMALIA-VL-SFT. Each subset is one
source dataset in the mix, each with a single train split. The only datasets that are absent from this mix are those that derive directly from the core LLM training mix, and can be found in the AMALIA-LLM Post Training Collection.
Example usage:
from datasets import… See the full description on the dataset page: https://huggingface.co/datasets/amalia-llm/AMALIA-VL-SFT-Dataset.llm_plan_gen_dataset_accu_t3
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t3.llm_plan_gen_dataset_accu_t1_t3
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t1_t3.jabarti-llm-dataset
jabarti-llm-dataset
Cleaned, section-chunked training corpus for a small bilingual LLM
(Arabic + English), combining a curated Egyptian-history collection with
general Wikipedia coverage from
CohereLabs/wikipedia-2023-11-embed-multilingual-v3.
Every pretrain record is a contiguous span of 120-1500 characters with the
article title and section headings removed. Provenance is in ds_source.
Configs and Splits
Config
Split
Rows
Training phase
Purpose… See the full description on the dataset page: https://huggingface.co/datasets/bakrianoo/jabarti-llm-dataset.llm_plan_gen_dataset_accu_t1
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t1.llm_plan_gen_dataset_accu_t2
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_accu_t2.llm_plan_gen_dataset_t0_t5
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_t0_t5.llm-math-evaluation-dataset
LLM Math Response Evaluation Dataset
Dataset Summary
A human-annotated dataset of 150 AI-generated math responses
evaluated across GPT-4o, Claude, and Gemini. Each response is
scored on Correctness, Reasoning, and Clarity using a structured
rubric, with written justification for every score.
Supported Tasks
LLM evaluation and benchmarking
Math reasoning quality assessment
Error type classification in AI responses
Dataset Structure… See the full description on the dataset page: https://huggingface.co/datasets/nbvbharath-1729/llm-math-evaluation-dataset.pde-llm-eval-code-perturbation-dataset
pde-llm-eval-code-perturbation-dataset
Code Perturbation Dataset (final). 256 PDE solver implementations: 64 base programs (32 physically valid, 32 with an injected bug that still runs to completion) expanded by 4 lexical perturbation conditions each -- original comments, comments removed, comments swapped in from a different implementation, and descriptive identifiers replaced by meaningless placeholders. The perturbations change the lexical surface only; executable behaviour… See the full description on the dataset page: https://huggingface.co/datasets/bermaneh/pde-llm-eval-code-perturbation-dataset.Korean-YouTube-Comment-Sentiment-Dataset
Korean YouTube Comment Sentiment Dataset
Data Overview
Summary
본 데이터셋은 유튜브에서 수집된 한국어 댓글 5,482개와 이에 대응하는 감정 레이블(긍정, 부정, 중립, 불명확)로 구성된 감정 분류용 데이터셋입니다.
주요 레이블: 긍정, 부정, 중립, 불명확
Features
수집 대상: 요리, 뷰티, 게임, 여행, 쇼핑 등 분야의 10만 명 이상 구독자를 보유한 유튜브 채널
형식: JSON (id, text, label)
검수: 한국인 검수자에 의한 수작업 라벨링 및 교차 검토
본 데이터셋은 구어체, 이모지, 줄임말 등 실제 사용자 표현이 반영되어 있습니다.
Dataset Structure
Dataset Fields
Field
Type
Description
id
string
각 댓글의… See the full description on the dataset page: https://huggingface.co/datasets/LLM-SocialMedia/Korean-YouTube-Comment-Sentiment-Dataset.mm-llm-coder-agent-dataset
Coder Agent Dataset
Agent workflow dataset for training coding agents. Contains multi-step coding tasks with tool usage patterns, execution validation, and quality metrics.
Skill Type: Agent/ Skill
This dataset is part of the combined Myanmar LLM dataset collection:
chat-skill.md - [amkyawdev/ myanmar-llm-data](https://huggingface. co/datasets/ amkyawdev/ myanmar-llm-data)
agent-skill.md - Myanmar conversational data, translations, Q&A
code-skill.md- [amkyawdev/… See the full description on the dataset page: https://huggingface.co/datasets/amkyawdev/mm-llm-coder-agent-dataset.ILSA-LLM-Extractor-Dataset
ILSA LLM Extractor Dataset
Project website: https://dedemerve.github.io/ILSA-LLM-Extractor/
Dataset Description
This dataset contains structured metadata automatically extracted from 1,756 peer-reviewed articles and reports covering International Large-Scale Assessments (IEA: TIMSS, PIRLS, ICCS; OECD: PISA, TALIS, PIAAC). The extraction pipeline combines PDF parsing, LLM-based structured extraction, and RAG-based synthesis.
Pipeline stages:
Stage 1: LLM-based… See the full description on the dataset page: https://huggingface.co/datasets/dedemerve/ILSA-LLM-Extractor-Dataset.Sensor_Driven_Environmental_Monitoring_LLM_Evaluation_Dataset
📌 Dataset Contents
Each sample includes:
category: The evaluation domain
prompt: The question given to the LLM
temperature: Environmental temperature input
humidity: Environmental humidity input
context: A scenario label (e.g., cool_humid, hot_dry, average_day)
reference: Expert-crafted expected output
All data is provided in a single JSON file.
🧪 Intended Use
This dataset supports research on:
LLM evaluation methods (semantic similarity, contextual… See the full description on the dataset page: https://huggingface.co/datasets/wayne-redemption/Sensor_Driven_Environmental_Monitoring_LLM_Evaluation_Dataset.llm_plan_gen_dataset_t0
[!IMPORTANT]
This is the training dataset for the ICAPS 2025 paper "Chasing Progress, Not Perfection: Revisiting Strategies for End-to-End LLM Plan Generation".
from pathlib import Path
import os
import jsonlines
from copy import deepcopy
from datasets import load_dataset
from icecream import ic
import enum
from enum import IntEnum
from enum import auto
class CONFIG_TYPES(enum.Enum):
# "type_id": ['t0', 'accu-t1', 'accu-t2', 'accu-t3', 'accu-t4', 'accu-t4', 'accu-t1+t4', 'accu-t2+t4']… See the full description on the dataset page: https://huggingface.co/datasets/huangsukai/llm_plan_gen_dataset_t0.DPO-Dataset
AMALIA DPO Dataset
This is the DPO (preference optimization) dataset used to train AMALIA-DPO.
It is a mix of preference pairs, mainly in European Portuguese and English, covering general conversation, instruction following, math, and safety. These pairs come from different sources, including prompts from the SFT mix, responses generated by different models, including an early version of the model, and some public datasets.
This dataset is provided as part of the AMALIA project… See the full description on the dataset page: https://huggingface.co/datasets/amalia-llm/DPO-Dataset.llm_guard_datasetstep2-evaluated-dataset-test2
Complete Evaluation Dataset (Rubric + LogP)
This dataset contains chain-of-thought explanations evaluated using both comprehensive rubric assessment and LogP evaluation.
Overview
Source Dataset: llm-compe-2025-kato/step2-evaluated-dataset-test2
Total Samples: 92
Successfully Evaluated (Rubric): 92
Failed Evaluations (Rubric): 0
Evaluation Model: Qwen/Qwen3-32B
Rubric Evaluation Results
Average Rubric Scores (0-4 scale)
logical_coherence: 3.51… See the full description on the dataset page: https://huggingface.co/datasets/llm-compe-2025-kato/step2-evaluated-dataset-test2.LLM-of-Babel-Final-Dataset-elstep2-evaluated-dataset-Qwen3-14B-cp32
Complete Evaluation Dataset (Rubric + LogP)
This dataset contains chain-of-thought explanations evaluated using both comprehensive rubric assessment and LogP evaluation.
Overview
Source Dataset: llm-compe-2025-kato/step2-evaluated-dataset-Qwen3-14B-cp32
Total Samples: 60
Successfully Evaluated (Rubric): 53
Failed Evaluations (Rubric): 7
Evaluation Model: Qwen/Qwen3-32B
Rubric Evaluation Results
Average Rubric Scores (0-4 scale)
logical_coherence:… See the full description on the dataset page: https://huggingface.co/datasets/llm-compe-2025-kato/step2-evaluated-dataset-Qwen3-14B-cp32.step2-evaluated-dataset-Qwen3-14B-cp40
Complete Evaluation Dataset (Rubric + LogP)
This dataset contains chain-of-thought explanations evaluated using both comprehensive rubric assessment and LogP evaluation.
Overview
Source Dataset: llm-compe-2025-kato/step2-evaluated-dataset-Qwen3-14B-cp40
Total Samples: 58
Successfully Evaluated (Rubric): 53
Failed Evaluations (Rubric): 5
Evaluation Model: Qwen/Qwen3-32B
Rubric Evaluation Results
Average Rubric Scores (0-4 scale)
logical_coherence:… See the full description on the dataset page: https://huggingface.co/datasets/llm-compe-2025-kato/step2-evaluated-dataset-Qwen3-14B-cp40.step2-evaluated-dataset-Qwen3-14B
Complete Evaluation Dataset (Rubric + LogP)
This dataset contains chain-of-thought explanations evaluated using both comprehensive rubric assessment and LogP evaluation.
Overview
Source Dataset: llm-compe-2025-kato/step2-evaluated-dataset-Qwen3-14B
Total Samples: 156
Successfully Evaluated (Rubric): 135
Failed Evaluations (Rubric): 21
Evaluation Model: Qwen/Qwen3-32B
Rubric Evaluation Results
Average Rubric Scores (0-4 scale)
logical_coherence:… See the full description on the dataset page: https://huggingface.co/datasets/llm-compe-2025-kato/step2-evaluated-dataset-Qwen3-14B.augmented_dataset_llm_generated_NER
📚 Augmented LLM-Generated NER Dataset for Scholarly Text
🧠 Dataset Summary
This dataset contains synthetically generated academic text tailored for Named Entity Recognition (NER) in the software engineering domain. The synthetic data augments scholarly writing using large language models (LLMs), with entity consistency maintained via token preservation.
The dataset is generated by merging and rephrasing pairs of annotated sentences from scholarly papers using… See the full description on the dataset page: https://huggingface.co/datasets/psresearch/augmented_dataset_llm_generated_NER.maze-llm-dataset
Multi-turn Maze Navigation Dataset
A dataset of maze navigation sequences for training AI agents to solve mazes through sequential decision-making.
Dataset Description
This dataset contains unique mazes with complete solution sequences. Each example includes:
A maze with visual path tracking
Sequential moves from start to goal
Minimum 2 moves per sequence
At least 2 different directions required per maze
Dataset Stats
Train Set
Total Sequences: 100… See the full description on the dataset page: https://huggingface.co/datasets/erdem-erdem/maze-llm-dataset.maze-llm-dataset-100K
Dataset Stats
Total Sequences: 106,436
Total Unique Mazes: 106,436
Total Moves: 1,648,442
Avg Moves per Maze: 15.5
Curriculum Distribution:
3x3: 6,436 mazes (6.0%), 45,858 moves
4x4: 20,000 mazes (18.8%), 197,308 moves
5x5: 20,000 mazes (18.8%), 255,138 moves
6x6: 20,000 mazes (18.8%), 318,744 moves
7x7: 20,000 mazes (18.8%), 381,744 moves
8x8: 20,000 mazes (18.8%), 449,650 moves
Dataset Quality Analysis:
Action distribution: {'N': 414286, 'S': 410290, 'E':… See the full description on the dataset page: https://huggingface.co/datasets/erdem-erdem/maze-llm-dataset-100K.