datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
data-pipeline-repair-trajectories
Data Pipeline Repair Trajectories
Rights & intended use: legacy public research corpus / portfolio
artifact. Hosted frontier-model outputs are research-only inputs under
project policy (synthetic-factory#161):
intended_use: research_only, project_training_policy: blocked. Not
training data for any model-weight update. Machine-readable record:
rights.json.
Release status: The raw, uncurated payload is now published under
data/raw/. It is available for inspection and… See the full description on the dataset page: https://huggingface.co/datasets/rmems/data-pipeline-repair-trajectories.SWEUniverse-Repaired-Indist-full-not-SWE-bench-pro-matched
VmaxRL/SWEUniverse-Repaired-Indist-full-not-SWE-bench-pro-matched
This dataset contains a 350-row subset selected from the Indist SWEUniverse training rows.
Selection policy: three-way repo overlap with Bugpilot and LM-Modify, deduped by repo plus introduction patch, then balanced round-robin across overlapping repos.
Rows: 350
Selected repos: 19
Deduped overlap capacity: 468
Source dataset: VmaxRL/SWEUniverse-Repaired-Indist-full-not-SWE-bench-pro-matched
repairllama-datasets
RepairLLaMA - Datasets
Contains the processed fine-tuning datasets for RepairLLaMA.
Instructions to explore the dataset
To load the dataset, you must define which revision (i.e., which input/output representation pair) you want to load.
from datasets import load_dataset
# Load ir1xor1
dataset = load_dataset("ASSERT-KTH/repairllama-datasets", "ir1xor1")
# Load irXxorY
dataset = load_dataset("ASSERT-KTH/repairllama-datasets", "irXxorY")
Citation
If you use… See the full description on the dataset page: https://huggingface.co/datasets/ASSERT-KTH/repairllama-datasets.VeriLoop-Structural-Repair-Verified
VLR-StructuralRepair v1.0.0 — non-regressive repair of real semantic defects
Evidence-convergent supervision for function-level semantic repair under a
hidden set of protected obligations. A candidate is positive only when it
preserves every already-satisfied obligation and strictly repairs at least
one. Aggregate improvement that breaks a protected obligation is a negative,
however far the total failure count drops.
The previous generation of this dataset… See the full description on the dataset page: https://huggingface.co/datasets/tsinghua-sigs-robot-lab/VeriLoop-Structural-Repair-Verified.Minecraft-GLB2Schem-RepairPairs-v1
unfundedResearcher/Minecraft-GLB2Schem-RepairPairs-v1
Paired (generated input, ground-truth target) Minecraft schematics for training a
model that turns an approximate voxelisation into a real build.
What a sample is
Each sample is three files inside a WebDataset TAR shard:
File
Meaning
<id>.input.schem
GENERATED. Produced by voxelising the source .glb. Approximate and noisy.
<id>.target.schem
GROUND TRUTH. The original schematic, copied byte-for-byte… See the full description on the dataset page: https://huggingface.co/datasets/unfundedResearcher/Minecraft-GLB2Schem-RepairPairs-v1.ReAPR-Automatic-Program-Repair-via-Retrieval-Augmented-Large-Language-ModelsThis is the Retrieval dataset used in the paper "ReAPR: Automatic Program Repair via Retrieval-Augmented Large Language Models"
dfm11-toolace-native-tool-use-repaired
dfm11-toolace-native-tool-use-repaired
ToolACE conversations with declared-name parsing and complete parallel result binding.
This is a DFM11 replacement for schneiderkamplab/dfm10-toolace-native-tool-use. All rows pass exhaustive structural validation. See metadata/manifest.json.
semantic-repair-routing
semantic-repair-routing
The 84,819 supervised pairs that trained
SemanticRepair-270M:
a message somebody actually wrote, and the requests inside it restated
plainly, one per line.
It teaches one narrow thing. An embedding router compares a question with
the description of every capability it can reach. People do not write the
way capabilities are described — they hedge, they apologise, they ask two
things in one breath, they name what they do not want. This data pairs
the first… See the full description on the dataset page: https://huggingface.co/datasets/Gramscii-IT/semantic-repair-routing.SWE-Repair
Dataset Summary
SWE-Repair is a curated subset of SWE-Bench, containing 204 single-function Python bugs from real-world GitHub repositories. Each example includes a buggy implementation and its corresponding problem statement.
Supported Tasks
Program Repair: Fixing bugs in Python functions
Code Generation: Generating correct implementations from buggy code
Dataset Structure
Each row contains:
instance_id: Unique identifier for the task (in format:… See the full description on the dataset page: https://huggingface.co/datasets/barty/SWE-Repair.code-contract-repaircode-contract-repair
APIContractRepair
APIContractRepair is a provenance-tracked instruction-tuning dataset for software engineers and code-model researchers who need contract-faithful, minimal repairs with tests that distinguish a broken implementation from its fix. Magicoder-OSS-Instruct-75K supplies real function-identifier seeds, but it does not provide these documented contracts, deliberately buggy implementations, minimal corrected implementations, or paired regression tests. This release… See the full description on the dataset page: https://huggingface.co/datasets/skonml/code-contract-repair.dfm11-synthetic-native-tool-calling-repaired
dfm11-synthetic-native-tool-calling-repaired
DFM8 synthetic tool trajectories with compatibility normalization materialized in source data.
This is a DFM11 replacement for schneiderkamplab/dfm8-synthetic-native-tool-calling. All rows pass exhaustive structural validation. See metadata/manifest.json.
context-repair-benchmark
ThoughtDAG Context Repair Benchmark
What happens after one wrong assumption enters a long LLM conversation?
This dataset turns context editing into a measurable intervention. Each synthetic case starts with a clean fact, introduces a false update, lets the error propagate through one to three downstream turns, and then asks the same final question under five graph conditions:
clean
polluted
source_prune
subgraph_prune
recompute_descendants
The central question is not only… See the full description on the dataset page: https://huggingface.co/datasets/thoughtdag/context-repair-benchmark.db-migration-repair-trajectories
Db Migration Repair Trajectories
Rights & intended use: legacy public research corpus / portfolio
artifact. Hosted frontier-model outputs are research-only inputs under
project policy (synthetic-factory#161):
intended_use: research_only, project_training_policy: blocked. Not
training data for any model-weight update. Machine-readable record:
rights.json.
Release status: The raw, uncurated payload is now published under
data/raw/. It is available for inspection and… See the full description on the dataset page: https://huggingface.co/datasets/rmems/db-migration-repair-trajectories.oellm-eu-defect-repair-sft-v1
oellm-eu-defect-repair-sft-v1
Monolingual SFT repair data for European-language generation defects observed
after Qwen 2B/4B/9B post-training. This dataset is designed to repair
degeneration, repetition loops, short answers, morphology damage, and
Bulgarian/Russian language leakage.
Strict SFT Schema
Each row in data/*.jsonl uses exactly:
{"messages":[{"role":"user","content":"..."},{"role":"assistant","content":"..."}],"lang":"is"}
Provenance, source URL… See the full description on the dataset page: https://huggingface.co/datasets/birgermoell/oellm-eu-defect-repair-sft-v1.OfficeSmith-PPTX-Repair
OfficeSmith PPTX Repair
Deterministically degraded PPTX IR objects paired with validated repairs.
Dataset summary
This dataset is part of the OfficeSmith collection for training models to plan, build, clarify, critique, and repair editable business presentations. It contains observable outputs only: no hidden chain of thought, secret benchmark prompt, personal data, or API credential is included.
Train rows: 160
Validation rows: 0
Test rows: 0
Languages: French… See the full description on the dataset page: https://huggingface.co/datasets/Benitoow/OfficeSmith-PPTX-Repair.hedgehog-stopping-repair-r5
hedgehog-stopping-repair-r5
Hedgehog — stopping-repair round 5.
Contents
train.jsonl (1848 rows)
validation.jsonl (438 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Original content for the Hedgehog extraction model (Michael Anthony Falabella).
SysMLv2_Repair_with_SLMs
SysMLv2 Repair with SLMs
Dataset used in "Automated Semantic Fault Localization in SysML v2: A Human-in-the-Loop Framework Using Knowledge-Graph Augmented LLMs", presented at INCOSE International Symposium 2026.
Dataset Structure
This dataset provides two configurations:
default: Contains train/validation/test splits used for fine-tuning small models. Samples exceeding 2048 tokens have been removed.
full: Contains complete dataset
Task
Given SysML v2 code… See the full description on the dataset page: https://huggingface.co/datasets/rohhaiil/SysMLv2_Repair_with_SLMs.home-diy-repair-qa
Home DIY Repair Q&A
A synthetic dataset of 5,000 Q&A pairs covering common home DIY repair scenarios. Each example includes a detailed step-by-step answer, required tools, safety warnings, and practical tips.
Dataset Purpose
This dataset is built for:
Instruction fine-tuning — train language models to give detailed, safe, and actionable home repair guidance
Retrieval-Augmented Generation (RAG) — build a knowledge base for home repair assistants
Question answering — train… See the full description on the dataset page: https://huggingface.co/datasets/dipenbhuva/home-diy-repair-qa.tb21-eval-qwen35-action-only-20k-infra-repaired-c164-max32k-timeout2x
qwen35-action-only-20k — Terminal-Bench 2.1
Noncanonical Terminal-Bench 2.1 evaluation of violetxi/qwen35-4b-offline-echo-action-only-20k-tacc through the served
model ID qwen35-action-only-20k with Terminus-2.
Noncanonical run: timeout_multiplier=2 instead of 1.0; repair concurrency=164 exceeds 30. Do not compare this score directly with canonical TB2.1 leaderboard runs.
Result
Recorded trials: 445
Tasks / attempts: 89 × 5
Errored trials scored as zero: 250… See the full description on the dataset page: https://huggingface.co/datasets/violetxi/tb21-eval-qwen35-action-only-20k-infra-repaired-c164-max32k-timeout2x.hedgehog-precision-repair
hedgehog-precision-repair
Hedgehog — precision-repair round (complete merchant extraction).
Contents
train.jsonl (1180 rows)
validation.jsonl (116 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Original content for the Hedgehog extraction model (Michael Anthony Falabella).
repa-fqt-sr-125k-fixture
REPA 125K stochastic-rounding validation fixture
This repository preserves the immutable inputs used to validate stochastic-rounding expectation for
cjf00000/REPA at a real terminal step-125000 checkpoint and rank-0 B64 training batch.
Authoritative artifacts
File
Bytes
SHA-256
checkpoint/0125000.pt
2,091,456,765
1c4df55bbe3ba7e2487b2af4f8ed4c6efb09e25740c902a27a9fb5417df8c19f
fixture/block6_125k_real_batch_sr_test_data.pt
466,343,899… See the full description on the dataset page: https://huggingface.co/datasets/jianfeichen/repa-fqt-sr-125k-fixture.hedgehog-complex-repair
hedgehog-complex-repair
Hedgehog — complex-extraction repair round.
Contents
train.jsonl (3840 rows)
validation.jsonl (304 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Original content for the Hedgehog extraction model (Michael Anthony Falabella).
douvras-lean-proof-repair
Douvras Lean Proof Repair Corpus
Exemplos sintéticos de erros comuns de reparo em Lean: importação ausente, incompatibilidade de
tipos, falha de tática, meta não resolvida, reescrita inválida e prova reflexiva. Os snippets não
foram executados no compilador (proof_status: NOT_EXECUTED); portanto o corpus não prova nenhum
teorema e não substitui validação com uma versão específica do Mathlib.
hedgehog-complex-implicit-repair
hedgehog-complex-implicit-repair
Hedgehog — complex implicit-schema repair round.
Contents
train.jsonl (2120 rows)
validation.jsonl (244 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Original content for the Hedgehog extraction model (Michael Anthony Falabella).
state-right-to-repair-laws
State Right-to-Repair Laws: Coverage, Requirements, and Effective Dates
Canonical, always-current version: https://referencesource.org/state-right-to-repair-laws/
Machine-readable: https://referencesource.org/state-right-to-repair-laws/data.json — this mirror is a point-in-time copy.
Last verified: 2026-08-15
Stale after: 2026-11-13 (past this date, prefer the canonical copy —
it re-verifies on a cadence this snapshot does not)
Records: 6
Which US states have enacted… See the full description on the dataset page: https://huggingface.co/datasets/referencesource/state-right-to-repair-laws.scugnizz-agentic-repair-50k-v2
Scugnizz Agentic Repair 50k
Dataset sintetico bilanciato per correggere renderer, copia esatta e tool calling.
Train: 49500
Validation: 500
Categorie:
{
"renderer_weather": 3750,
"renderer_finance": 3750,
"renderer_spotify": 8,
"renderer_mail": 3750,
"renderer_calendar": 448,
"renderer_dns": 36,
"renderer_whois": 3750,
"renderer_json_complex": 3750,
"exact_hash": 64,
"exact_network": 180,
"exact_url_domain": 2424,
"tool_weather": 48,
"tool_finance": 36… See the full description on the dataset page: https://huggingface.co/datasets/ProjectScugnizz/scugnizz-agentic-repair-50k-v2.repairB0-repair-DPO
B0-repair-DPO
69 on-policy tool-choice preference pairs — the repair rung of
schneewolflabs/B0-9B.
Method: fresh scenarios (LLM-generated, 8-gram-filtered against the eval set) in the
four shapes the model failed — delegate-to-code_agent engineering problems, read_file
lookups, execute_command box-state questions, git_diff staged/unstaged. rejected is
the model's own real wrong response sampled under its serving system prompt; chosen
is the gold tool call. Scenarios the model… See the full description on the dataset page: https://huggingface.co/datasets/schneewolflabs/B0-repair-DPO.bibletts-asante-twi-repaired
BibleTTS Asante Twi — Repaired Transcripts
The Asante Twi transcripts released with BibleTTS have had the
characters ɛ (U+025B) and ɔ (U+0254) stripped out. This dataset restores them.
Audio is not included. This is a drop-in replacement for the .txt files that ship with the
BibleTTS Asante Twi package, matched by clip ID.
The problem
Both are Twi vowels, and both are required by the orthography. Measured across the released
Asante Twi transcripts:
Character… See the full description on the dataset page: https://huggingface.co/datasets/danieldzikunuofmarvel/bibletts-asante-twi-repaired.