Browse Articles

Large language models are increasingly used to solve math problems, but their ability to handle multi-step reasoning remains uncertain. In this study, students tested whether different prompting styles could improve LLM accuracy on challenging AIME math questions and found that detailed step-by-step solutions did not significantly outperform simpler prompts. These results suggest that improving LLM mathematical reasoning may require deeper model-level advances rather than changes in prompting style alone.

What type of motivation is more effective: reward or punishment? In this study, the authors assess the effects of positive or negative on the math scores of sixth graders.

In this study, the authors address potential reasons why employees may voluntarily resign. This is in response to the currently observed economic trend The Great Resignation. Through analysis of federal and local government data along with survey results from Fairfax County, they concluded that adding additional benefits will help companies retain talented empolyees.