Crossing the NL/PL Divide: Information Flow Analysis Across the NL/PL Boundary in LLM-Integrated Code

1. 1Identify LLM API calls in the code
2. 2Analyze the natural-language prompts and their impact on the program's behavior
3. 3Develop a method to track data flow across the NL/PL boundary
4. 4Apply the method to existing program analysis techniques, such as taint analysis and program slicing

This research benefits software engineers and AI researchers working on LLM-integrated systems, as it enables them to better understand and analyze the information flow across the NL/PL boundary

💡 Existing program analysis techniques are limited by their inability to cross the NL/PL boundary, hindering their effectiveness in LLM-integrated systems

💡 Analyzing info flow across NL/PL boundary in LLM-integrated code