Dimensions overall score 7.0

GitHub Code Pulse

No public code linked for this paper yet.

• Leading Large Language Models (LLMs) achieve a higher F1 score (0.96) than traditional static analyzers and machine learning models (up to 0.87) on an aggregated benchmark of real-world contracts in a zero-shot setting.result 95%
• The Reentrancy Scenarios Dataset (RSD) was constructed with 143 handcrafted minimal working examples to isolate and stress-test individual reentrancy patterns.method 95%
• Most reentrancy detection tools fail on multiple scenarios within the Reentrancy Scenarios Dataset (RSD).limitation 90%
• The best performing model on the Reentrancy Scenarios Dataset (RSD) attains an F1 score of 0.82.result 95%
• Existing datasets for reentrancy detection often embed tool-induced inconsistencies and contradictory labels.limitation 90%
• The paper evaluates 12 formal-methods-based tools, 10 machine-learning models, and 9 large language models for reentrancy detection.method 95%
• The Aggregated Benchmark consists of 432 real-world contracts consolidated and relabeled from prior datasets.method 95%
• Leading Large Language Models (LLMs) achieve a higher F1 score (0.96) than traditional static analyzers and machine learning models (up to 0.87) on a benchmark of real-world contracts in a zero-shot setting.result 95%
• The Reentrancy Scenarios Dataset (RSD) was constructed with 143 handcrafted minimal working examples to isolate and stress-test individual reentrancy patterns.method 95%
• Most reentrancy detection tools fail on multiple scenarios within the Reentrancy Scenarios Dataset (RSD).limitation 90%
• The Aggregated Benchmark consists of 432 real-world contracts, consolidated and relabeled from prior datasets.method 95%
• The study evaluates 12 formal-methods-based tools, 10 machine-learning models, and 9 large language models for reentrancy detection.method 95%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card