Can I understand what I create? Self-Knowledge Evaluation of Large Language Models
This work addresses the need for robust evaluation frameworks for LLMs, offering an easy-to-implement method that could improve model assessment and development, though it appears incremental in nature.
The authors tackled the problem of evaluating large language models' self-knowledge by introducing a framework that tests models on their ability to comprehend and respond to self-generated questions, revealing significant gaps in this ability and showing that fine-tuning on self-generated math tasks can enhance math performance.
Large language models (LLMs) have achieved remarkable progress in linguistic tasks, necessitating robust evaluation frameworks to understand their capabilities and limitations. Inspired by Feynman's principle of understanding through creation, we introduce a self-knowledge evaluation framework that is easy to implement, evaluating models on their ability to comprehend and respond to self-generated questions. Our findings, based on testing multiple models across diverse tasks, reveal significant gaps in the model's self-knowledge ability. Further analysis indicates these gaps may be due to misalignment with human attention mechanisms. Additionally, fine-tuning on self-generated math task may enhance the model's math performance, highlighting the potential of the framework for efficient and insightful model evaluation and may also contribute to the improvement of LLMs.