Mothman at SemEval-2024 Course of 9: An Iterative System for Chain-of-Thought Instant Optimization
Authors: Alvin Po-Chun Chen, Ray Groshan, Sean von Bayern
Abstract: Intensive evaluation exists on the effectivity of giant language fashions on logic-based duties, whereas comparatively little has been accomplished on their potential to generate inventive choices on lateral pondering duties. The BrainTeaser shared exercise assessments lateral pondering and makes use of adversarial datasets to forestall memorization, resulting in poor effectivity for out-of-the-box fashions. We propose a system for iterative, chain-of-thought fast engineering which optimizes prompts using human evaluation. Using this shared exercise, we exhibit our system’s potential to significantly improve model effectivity by optimizing prompts and take into account the enter dataset
