zkaedi/gemma-2-9b-solidity-merged
The zkaedi/gemma-2-9b-solidity-merged model is a 9.2 billion parameter language model based on Google's Gemma-2-9B architecture. It has been fine-tuned with a LoRA adapter on 2,250 Solidity vulnerability energy signatures, covering 19 distinct classes. This specialized model is optimized for identifying and understanding vulnerabilities within Solidity smart contracts, making it highly effective for blockchain security analysis.
Loading preview...
Model Overview
zkaedi/gemma-2-9b-solidity-merged is a specialized language model built upon Google's Gemma-2-9B architecture, featuring 9.2 billion parameters and a 16384-token context length. Its core differentiator is the integration of a LoRA adapter, which was fine-tuned using a dataset of 2,250 Solidity vulnerability energy signatures across 19 distinct classes. This fine-tuning process, conducted via SFT using Unsloth/TRL, has permanently merged the adapter into the base model.
Key Capabilities
- Solidity Vulnerability Detection: Enhanced understanding and identification of common and complex vulnerabilities in Solidity smart contracts.
- Specialized Knowledge: Deep expertise in 19 specific categories of Solidity security flaws, derived from energy signatures.
- Gemma-2-9B Foundation: Benefits from the robust language understanding and generation capabilities of the underlying Gemma-2-9B model.
Good For
- Smart Contract Security Audits: Assisting in the automated or semi-automated analysis of Solidity code for security weaknesses.
- Blockchain Development: Providing insights into potential vulnerabilities during the development phase of decentralized applications (dApps).
- Security Research: Aiding researchers in exploring and categorizing Solidity smart contract vulnerabilities.