Add to Favourites
To login click here
This article discusses the limitations of large language models (LLMs) in tasks that require visual and spatial reasoning and explores various approaches to improve their performance. These include chain-of-thought prompting, tool usage and code augmentation, and the use of whiteboards to enhance the visual reasoning abilities of multimodal large language models (MLLMs). Researchers from Columbia University have proposed a new approach called Whiteboard-of-Thought (WoT) prompting, which allows MLLMs to draw out reasoning steps as images.
