The hyperparameters for training are detailed in Table 7. We perform the human evaluation on two of the four in-depth knowledge quality assessment metrics. V alidity ( "): whether the generated visual knowledge is valid to humans . Conformity ( "): whether the generated knowledge faithfully depicts the scenarios in the images . Our calculated average pairwise Cohen's Suppose you are looking at an image that contains the following subject and object entities: Subject list: [Insert the subject names here] Object list: [Insert the object names here] Please extract 5-10 condensed descriptions that describe the interactions and/or relations among those entities in the image.

The use of large language models (LLMs) in natural language processing (NLP) tasks is rapidly increasing, leading to changes in how researchers approach problems in the field. To fully utilize these models' abilities, a better understanding of their behavior for different input protocols is required. With LLMs, users can directly interact with the models through a text-based interface to define and solve various tasks. Hence, understanding the conversational abilities of these LLMs, which may not have been specifically trained for dialog modeling, is also important. This study examines different approaches for building dialog systems using LLMs by considering various aspects of the prompt. As part of prompt tuning, we experiment with various ways of providing instructions, exemplars, current query and additional context. The research also analyzes the representations of dialog history that have the optimal usable-information density. Based on the findings, the paper suggests more compact ways of providing dialog history information while ensuring good performance and reducing model's inference-API costs. The research contributes to a better understanding of how LLMs can be effectively used for building interactive systems.