Yunfeng Shi

TurtleBench: Evaluating Top Language Models via Real-World Yes/No Puzzles

As the application of Large Language Models (LLMs) expands, the demand for reliable evaluations increases. Existing LLM evaluation benchmarks primarily rely on static datasets, making it challenging to assess model performance in dynamic interactions with users. Moreover, these benchmarks often depe…