datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
CS605-NLP
Automating Legal Defense: Strategy Planning with Large Language Models
📖 Overview
This repository contains the code, data, and models for an end-to-end AI framework that models defense strategising in Singapore criminal court cases as a sequential decision task. The project was developed as part of CS605 - Natural Language Processing course at the Master of IT in Business (MITB) program, Singapore Management University.
Authors
WANG Gan
CHUA Xin Yu… See the full description on the dataset page: https://huggingface.co/datasets/Errorman23/CS605-NLP.albanian-error-augmentation
Albanian Controlled Error Augmentation Dataset
Dataset of controlled Albanian orthographic errors created for PhD research on Albanian spelling education and automatic exercise generation.
Each row is an (incorrect → correct) pair with an explicit error_type label.
Error types
error_type
Description
missing_diacritic
Missing ë / ç
c_q_confusion
Confusion between ç / q / c
digraph_reduction
Digraph loss (sh, dh, th, gj, nj, ll, rr, xh, zh)… See the full description on the dataset page: https://huggingface.co/datasets/greta44/albanian-error-augmentation.mondegreen-asr-errors
Mondegreen ASR error pairs
(ASR hypothesis, gold text) pairs for Japanese ASR post-correction.
This build is simulated -- errors come from a phonetic corruption model, not from a real ASR system. It exists so the whole pipeline (gate training, benchmarks, figures, CI) is reproducible without a GPU. Treat every number derived from it as a stated assumption, not a measurement.
How it was made
synthetic text
-> phonetic corruption model (mondegreen.simulate)
->… See the full description on the dataset page: https://huggingface.co/datasets/NagaYu/mondegreen-asr-errors.synthetic-real-word-errors
Synthetic Real-Word Error Datasets
This repository contains synthetic German data for grammatical error detection and correction, with a focus on context-dependent real-word errors.
The repository provides four subsets:
Subset
Description
Examples
mixed_real_word
Mixed real-word errors
99,812
capitalization
Capitalization errors
99,664
case
Case errors
99,706
verb
Verb errors
99,780
Each subset contains both erroneous and correct sentences and can therefore… See the full description on the dataset page: https://huggingface.co/datasets/aurorra/synthetic-real-word-errors.quantum-error-mitigation-and-benchmarking
Neura Parse — Quantum Error Mitigation, Characterization & Benchmarking
A pre-fault-tolerance, code-backed vertical on getting trustworthy answers from noisy hardware and rigorously measuring device quality: error-mitigation techniques, characterization/tomography protocols, and benchmarking suites. Runnable Mitiq, pyGSTi, and Qiskit Experiments pipelines with honest sampling-overhead and bias/variance accounting — the practitioner and research toolkit the general dataset… See the full description on the dataset page: https://huggingface.co/datasets/Neura-parse/quantum-error-mitigation-and-benchmarking.omnimcp_type_error_mypy_resolver_teaser
🔬 INSPECT THE DEEPSEEK-R1 REASONING CHAIN LIVE:
Zero hallucinations. Null syntax errors. 100% AST compiler validated.🌐 Live Interactive Reasoning & Code Inspector: https://emgena.com/trainingslager🎁 Claim your Free Starter Kit (Code: STARTER100): https://emgena.com/trainingslager🏷️ Launch Discount: Get 20 € OFF any 500-incident production suite with code LAUNCH20!
📜 Enterprise Compliance: EU AI Act Articles 50 & 53 certified • 100% DSGVO / GDPR clean • Commercial EULA… See the full description on the dataset page: https://huggingface.co/datasets/emgena/omnimcp_type_error_mypy_resolver_teaser.python-runtime-verified-error-correction
Python Runtime-Verified Error Correction Dataset 🐍⚡
Overview
Production-grade synthetic dataset of Python code errors with runtime-verified corrections. Each sample contains broken code, the actual runtime error, and a guaranteed-working fix validated through execution.
Unlike traditional synthetic datasets, every correction is verified by actually running the code in an isolated environment—eliminating hallucinations and ensuring real-world applicability.… See the full description on the dataset page: https://huggingface.co/datasets/SyntheticLogic-Labs/python-runtime-verified-error-correction.arabic-grammar-errorsErrorBench
ErrorBench: Fine-Grained Error Analysis of Multi-Family LLMs in Data-to-Text Generation
Dataset Summary
ErrorBench is a human-annotated, span-level benchmark for analyzing generation errors in Large Language Models (LLMs) for Data-to-Text (D2T) generation. The dataset consists of sentences generated from structured DBpedia triples and annotated with fine-grained span-level error labels across 10 error categories.
The dataset was introduced in our IJCNN 2026 paper… See the full description on the dataset page: https://huggingface.co/datasets/soumyaBharadwaj/ErrorBench.errored_pythonThis is a subset of the python dataset provided but Ailurophile on Kaggle.
Important:Errors were introduced on purpose to try to test a sort of "specialized masking" in a realistic way.
Goal:The goal is to create a specialized agent, and add it to a chain with at least one other agent that generates code, and can hopefully "catch" any errors.
Inspiration:When working to generate datasets with other models, I found that even after multiple "passes" errors where still missed.
Out of curiosity… See the full description on the dataset page: https://huggingface.co/datasets/TacoPrime/errored_python.openslr-sinhala-synthetic-spell-errors-quarter
Sinhala Dyslexic Spelling Correction Dataset
Dataset Description
This dataset contains Sinhala and code-mixed (Sinhala-English) text pairs for training spelling correction models, specifically designed to address dyslexia-like spelling errors.
Features
dyslexic_sentence: Input text with dyslexia-like spelling errors (string)
correct_sentence: Corrected output text (string)
Dataset Statistics
Split
Samples
Train
37,056
Test
9,265… See the full description on the dataset page: https://huggingface.co/datasets/SPEAK-PP/openslr-sinhala-synthetic-spell-errors-quarter.Vietnamese_spelling_error
Vietnamese Spelling Error Dataset
This dataset contains examples of Vietnamese text with spelling errors and their corresponding corrections. It is intended to be used for training and evaluating models in spelling correction tasks, particularly for the Vietnamese language.
Dataset Summary
Name: Vietnamese Spelling Error Dataset
Language: Vietnamese
File Format: [CSV/Parquet/dataset/etc.]
Columns:
text: The corresponding corrected version of the text.
error_text: The… See the full description on the dataset page: https://huggingface.co/datasets/ShynBui/Vietnamese_spelling_error.codenet-compile-errors
CodeNet Compiler Errors (Re-compiled 2026)
Dataset Summary
This dataset contains source code submissions from Project CodeNet that fail to compile. Unlike the original dataset metadata (which reflects compiler versions from 2011–2020), this dataset was re-executed in a modern Debian environment (2026) to generate up-to-date compiler error messages.
It is designed for research in:
Automated Program Repair (APR): Fixing compile-time errors.
Compiler Error Explanation:… See the full description on the dataset page: https://huggingface.co/datasets/criyle/codenet-compile-errors.error-detection-positives
error-detection-positives
This dataset is part of the PARC (Premise-Annotated Reasoning Collection) and contains mathematical reasoning problems with error annotations. This dataset combines positives samples from multiple domains.
Domain Breakdown
gsm8k: 50 samples
math: 53 samples
metamathqa: 93 samples
orca_math: 96 samples
Features
Each example contains:
data_source: The domain/source of the problem (gsm8k, math, metamathqa, orca_math)
question: The… See the full description on the dataset page: https://huggingface.co/datasets/PARC-DATASETS/error-detection-positives.error-detection-negatives
error-detection-negatives
This dataset is part of the PARC (Premise-Annotated Reasoning Collection) and contains mathematical reasoning problems with error annotations. This dataset combines negatives samples from multiple domains.
Domain Breakdown
gsm8k: 57 samples
math: 44 samples
metamathqa: 59 samples
orca_math: 54 samples
Features
Each example contains:
data_source: The domain/source of the problem (gsm8k, math, metamathqa, orca_math)
question: The… See the full description on the dataset page: https://huggingface.co/datasets/PARC-DATASETS/error-detection-negatives.Referencing_Errors_Synthetic_EN
Synthetic Dataset for Automatic Error Correction in Referencing
This dataset includes 4,600 parallel sentences for Automatic Error Correction in referencing in German.It was synthetically created with gpt-4o-mini model according to the Institutional Guidelines of the Center for Translation Studies (CTS), University of Vienna.
Dataset Description
corrupted_sentence: the sentence containing the referencing error
clean_sentence: the correct version of the corrupted sentence… See the full description on the dataset page: https://huggingface.co/datasets/elizaveta-dev/Referencing_Errors_Synthetic_EN.python-errors-and-fixes-dataset
Python Error Fixing Dataset
📋 Overview
A curated dataset of 200 unique Python error scenarios designed for fine-tuning Large Language Models (LLMs) on debugging and error-fixing tasks. Each example simulates a real-world programming issue, complete with code, traceback, and a comprehensive solution.
Dataset Statistics
Total Examples: 200
Error Categories: 50+ distinct error types
Python Version: 3.12+
License: MIT
🎯 Purpose
This… See the full description on the dataset page: https://huggingface.co/datasets/salikahmad702/python-errors-and-fixes-dataset.error-detection-positives_perturbed
error-detection-positives_perturbed
This dataset is part of the PARC (Premise-Annotated Reasoning Collection) and contains mathematical reasoning problems with error annotations. This dataset combines positives_perturbed samples from multiple domains.
Domain Breakdown
gsm8k: 48 samples
math: 42 samples
metamathqa: 72 samples
orca_math: 85 samples
Features
Each example contains:
data_source: The domain/source of the problem (gsm8k, math, metamathqa… See the full description on the dataset page: https://huggingface.co/datasets/PARC-DATASETS/error-detection-positives_perturbed.ocr2_cf1900_k2_gpt55_medium_qwen35_error_steps_seed20260513
GPT-5.5 Medium Reannotation of Qwen3.5-Positive OCR2 Coding Steps
This dataset follows the same 500-row parquet layout as JingweiNi/ocr2_cf1900_k2_qwen35_fp8_10k_seed20260513 and contains GPT-5.5 medium-reasoning reannotations for the 1,536 Qwen3.5-positive error steps.
Summary
Source dataset: JingweiNi/ocr2_cf1900_k2_qwen35_fp8_10k_seed20260513
Source rows: 500 K2-Think Codeforces traces
Source manifest-selected Qwen3.5 labels: 10,000 steps
GPT-5.5 reannotated… See the full description on the dataset page: https://huggingface.co/datasets/JingweiNi/ocr2_cf1900_k2_gpt55_medium_qwen35_error_steps_seed20260513.DeepScaleR-Qwen3-1.7B-2k-strategy-error-200
DeepScaleR Qwen3 1.7B 2K strategy errors
This dataset contains 200 distinct questions selected from
zjhhhh/DeepScaleR-Qwen3-1.7B-2k-agreed-regraded-le5-coded
at revision 8b6e0f481bced00132c95fb631745d4992fa19fd.
Each row has one manually selected model response whose main failure is a
strategy error relative to the source row's code_hint: the response does not
materially use the hint's core route, substitutes another strategy, or omits a
decisive hinted stage in favor of an… See the full description on the dataset page: https://huggingface.co/datasets/zjhhhh/DeepScaleR-Qwen3-1.7B-2k-strategy-error-200.Errors_Mod This dataset is the result of errors found in generated ecore files by different LLMs, mainly GTP4-Turbo and Llama3-70b-Instruct.
The errors have been classified into :
Wrong Type : This can occur if the generated type is non existant or used in a wrong way
Missing declaration : this can be due to either a missing declaration like xsi or nonexistant one
Start Token : this can mostly be due to start tag <?xml ..> <ecore ..> that are missing, happens when we can't read the file or error in… See the full description on the dataset page: https://huggingface.co/datasets/VeryMadSoul/Errors_Mod.instruction-error-detection-en-id
instruction-error-detection-en-id
Description
instruction-error-detection-en-id is a bilingual benchmark dataset for detecting, explaining, and correcting flawed or ambiguous instructions.
The dataset focuses on instruction robustness by introducing graded difficulty levels and partially incorrect instructions. It is designed to evaluate how well models can reason about contradictions, ambiguities, and incomplete constraints before responding.
Dataset Fields… See the full description on the dataset page: https://huggingface.co/datasets/YosepMulia/instruction-error-detection-en-id.granite-base-model-errors
Granite-1B Base Model Errors
Overview
This dataset contains 10 examples where the Granite-4.0-1B-Base language model produces incorrect or awkward outputs. Each row includes:
id: a unique identifier for each example
input: the prompt given to the model
expected_output: what the correct answer or completion should be
model_output: what the model actually produced
The dataset demonstrates common blind spots of a base causal language model, including factual errors, logic… See the full description on the dataset page: https://huggingface.co/datasets/thatgirltomiie/granite-base-model-errors.AstralMath-v1-ErrorTracesChangelog:
2026-03-26:
Public AstralMath-v1-ErrorTraces, include 520k error traces that models encounter during the synthesis process.
2026-03-25:
Add new 50k datapoints, replace ~10k old datapoints with higher quality synthetic questions(12 consensus tranform use tool for verify) to stage 1.
Removed ~600 datapoints affect by extract function bug(raise incomplete question).
Replace 1 question in AstralBench(hmmt-feb-2026-algebra-p7 -> open-rl-combinatorics-247678).
2026-03-12:
Release… See the full description on the dataset page: https://huggingface.co/datasets/nguyen599/AstralMath-v1-ErrorTraces.robot-error-recovery-tr
Robot Error Recovery Dataset (TR)
This dataset teaches robots how to react when a task cannot be completed successfully.
Instead of normal navigation commands, this dataset focuses on failure situations and recovery behaviors.It is designed for embodied AI systems, service robots and home assistant robots.
Structure
Each entry contains:
situation: what went wrong in the environment
recovery_action: what the robot should do next
Example
situation: Robot cannot… See the full description on the dataset page: https://huggingface.co/datasets/vosap52/robot-error-recovery-tr.robot-error-correction-tr-v1
Robot Error Correction TR v1
This dataset focuses on failure detection and corrective behavior in embodied AI systems.
Unlike standard instruction datasets, each sample represents:
an incorrect real-world outcome
a corrective decision
The goal is improving humanoid robot autonomy and reliability in real environments.
Capabilities trained:
self-correction
safety awareness
environment feedback handling
recovery planning
Referencing_Errors_Synthetic_DE
Synthetic Dataset for Automatic Error Correction in Referencing
This dataset includes 4,600 parallel sentences for Automatic Error Correction in referencing in German.It was synthetically created with gpt-4o-mini model according to the Institutional Guidelines of the Center for Translation Studies (CTS), University of Vienna.
Dataset Description
corrupted_sentence: the sentence containing the referencing error
clean_sentence: the correct version of the corrupted sentence… See the full description on the dataset page: https://huggingface.co/datasets/elizaveta-dev/Referencing_Errors_Synthetic_DE.Funny-Windows-Errors-Windows-Bibles