datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
solidity-audit-cot
solidity-audit-cot
Long-CoT audit traces for Solidity contracts, generated by Claude Opus 4.7 (adaptive thinking, xhigh effort) over the spec→contract corpus from the Qwopus3.6-27B-solidity training pipeline.
This dataset is the Stage 2 training corpus for the multi-stage Qwopus3.6-27B-solidity model — designed to teach long-form security reasoning (8-15 paragraph chain-of-thought) anchored to real Solidity contracts.
Why this dataset exists
Public Solidity audit… See the full description on the dataset page: https://huggingface.co/datasets/samscrack/solidity-audit-cot.solidity_vulnerability_audit_dataset
Solidity Vulnerability Audit Dataset
Organization: gitmate AI
Dataset Summary
The Solidity Vulnerability Audit Dataset is a curated collection of Solidity smart contract code snippets paired with expert-written vulnerability audits. Each entry presents a real or realistic smart contract scenario, and the corresponding analysis identifies security vulnerabilities or confirms secure patterns. The dataset is designed for instruction-tuned large language models (LLMs) to… See the full description on the dataset page: https://huggingface.co/datasets/GitmateAI/solidity_vulnerability_audit_dataset.solidity-security-findings
Solidity Security Findings Dataset
9,359 smart contract security findings from Code4rena audits.
Statistics
HIGH severity: 4,418 findings
MEDIUM severity: 4,910 findings
Sources: Code4rena, Immunefi
Use Cases
Training security audit models
Vulnerability classification
Smart contract analysis
solidity-audit-instruct
solidity-audit-instruct
An instruction-tuning dataset for Solidity smart-contract security auditing,
in TRL SFT messages format. Built from
msc-smart-contract-auditing/audits-with-reasons
(config default, split train), shuffled with seed 42 and split 90/10.
Row counts
train.jsonl: 2224 rows
eval.jsonl: 248 rows
total: 2472 rows
Format
One JSON object per line (JSONL), TRL SFT messages format:
{"messages": [
{"role": "system", "content":… See the full description on the dataset page: https://huggingface.co/datasets/Offlin33er/solidity-audit-instruct.starcoder_solidity_finetuneSolidity-contract-Specification-dataset-updatedsolidity_vulnerability_audit_dataset
Solidity Vulnerability Audit Dataset
Organization: gitmate AI
Dataset Summary
The Solidity Vulnerability Audit Dataset is a curated collection of Solidity smart contract code snippets paired with expert-written vulnerability audits. Each entry presents a real or realistic smart contract scenario, and the corresponding analysis identifies security vulnerabilities or confirms secure patterns. The dataset is designed for instruction-tuned large language models (LLMs) to… See the full description on the dataset page: https://huggingface.co/datasets/xj210/solidity_vulnerability_audit_dataset.solidity-audit-instruct-v2
Solidity Audit Instruct V2 — Canonical Taxonomy
V2 of Offlin33er/solidity-audit-instruct with all
finding labels normalized into 8 canonical classes (Reentrancy, Arithmetic Error, Access Control
Flaw, Input Validation Error, Frontrunning / MEV, Denial of Service (DoS), Business Logic
Violation, Clean (No Vulnerability)).
Split-preserving: V1's train (2,224) and test (248) rows are transformed in place —
the held-out set contains identical code to V1's, so eval numbers stay… See the full description on the dataset page: https://huggingface.co/datasets/Offlin33er/solidity-audit-instruct-v2.Slither_Analyzed_Solidity_DatasetThis dataset consists of 40,272 Slither analyzed Solidity files. The files were gathered from three sources: Etherscan, Github and DISL dataset. Three preprocessing steps were applied:
Step 1 "Cleaning": Unnecessary parts such as comments or blank lines were removed from each file.
Step 2 "Formatting": Each file was converted with Prettier and the corresponding Solidity-plugin.
Step 3 "Slither Analysis": Each file has been checked for security vulnerabilities and gas consumption issues.… See the full description on the dataset page: https://huggingface.co/datasets/fbnhnsl/Slither_Analyzed_Solidity_Dataset.solidity_evaloptimized-solidity-datasetgemma_refine_soliditystarcoder_3b_baseline_soliditysolidity-expert-datasetgemma-starcoder-baseline-solidityPreprocessed_Solidity_Dataset_V1This dataset consists of 4,134 unique Solidity files. The files were gathered from three sources: Etherscan, Github and DISL dataset. Six preprocessing steps were applied:
Step 1 "Cleaning": Unnecessary parts such as comments or blank lines were removed from each file.
Step 2 "Formatting": Each file was converted with Prettier (and the corresponding Solidity-plugin) so that the final model only generates code in a correct format.
Step 3 "Slither Analysis": Each file has been checked for… See the full description on the dataset page: https://huggingface.co/datasets/fbnhnsl/Preprocessed_Solidity_Dataset_V1.solidity-fluency-datagemma-2b-baseline-solidityw3sa-bm-solidity
W3SA - Solidity Repo Benchmark
This benchmark evaluates audited EVM projects containing smart contracts written in Solidity. It’s designed to assess the vulnerability detection capabilities of AI models and static analysis tools. The benchmark tests the ability of different models to detect high and medium severity vulnerabilities. Our model ALMX 1.5 demonstrates a significant advantage by catching up to 5x more high-severity vulnerabilities than traditional static analysis tools… See the full description on the dataset page: https://huggingface.co/datasets/almanax/w3sa-bm-solidity.solidity_qa
solidity_qa
This dataset was uploaded automatically.