GPT-4

The 'GPT-4' dataset/benchmark is used to evaluate the effectiveness of context-based adversarial attacks on AI code generators, specifically analyzing their vulnerability to generating exploitable code under strategically crafted inputs.

Papers using GPT-4 (5)

Context-Based Adversarial Attacks on AI Code Generators: Vulnerability Analysis and Implications2026MASTEST: A LLM-Based Multi-Agent System For RESTful API Tests2025GenSim: Generating Robotic Simulation Tasks via Large Language Models2023 · 8 citesHire a Linguist!: Learning Endangered Languages with In-Context Linguistic Descriptions2024 · 1 citesData-Efficient Learning with Neural Programs2024