- Performed comprehensive data visualization to analyze performance accuracy of LLMs across different graph sizes using Python
- Developed a novel translator function for DFS within the LLM-CLRS Graph Reasoning Benchmark, transforming intermediate steps and predictions into a structured hint format, thereby enhancing the evaluation and training of LLMs in solving complex graph-based
- Developing a website - using React, TypeScript, and Node.js - for the LLM-CLRS Graph Reasoning Benchmark featuring integrated