Towards Harnessing Large Language Models for Enhanced Comprehension of Conversational Grounding

Overview

This repository contains the source code and data for the study "Towards Harnessing Large Language Models for Enhanced Comprehension of Conversational Grounding" published at IWSDS 2024.

Cite as: Kristiina Jokinen, Phillip Schneider, Taiga Mori (2024). Towards Harnessing Large Language Models for Enhanced Comprehension of Conversational Grounding. Proceedings of the 14th International Workshop on Spoken Dialogue Systems Technology (IWSDS 2024), Sapporo, Japan.

Contents

1. Notebook: The Jupyter notebook (conv-grounding-llm-prediction.ipynb) details the experimental procedures, prompts, and configuration of the large language model.

2. Datasets: The dataset annotated_dialogues.xlsx was utilized in the experiment as it contains the input dialogues. The LLM predictions are provided as four additional datasets.

How to Use

Jupyter Notebook

• Open the Jupyter notebook using a compatible environment or platform.
• Follow the step-by-step instructions to reproduce the experiment.