OAK Security SFT: Bilingual Instruction-Tuning Corpus for On-Chain Attack Analysis

A bilingual English and Russian instruction-tuning dataset for training language models to analyze on-chain attacks, DeFi exploits, and crypto security incidents. It was created by author z0n3x and is grounded in the OAK taxonomy v0.1, a structured knowledge base of adversary tactics, techniques, and real-world incidents.

Use Cases

• Instruction-tuning language models for blockchain forensics based on the described taxonomy of adversary tactics and techniques.
• Analyzing DeFi exploits and security incidents based on the structured knowledge base of real-world on-chain events.
• Training models for multilingual security analysis based on the bilingual English and Russian content.

Strengths

• Grounded in the structured OAK taxonomy v0.1, which includes adversary tactics, techniques, and mitigations.
• Contains bilingual English and Russian content for instruction-tuning.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Freshness

Last updated 2026-06-10 12:06:29; freshness should be verified.