Evaluating Linguistic Capabilities of Multimodal LLMs in the Lens of Few-Shot Learning

TL;DR

This paper evaluates the linguistic and reasoning capabilities of Multimodal Large Language Models (MLLMs) using the VALSE benchmark, highlighting the impact of few-shot learning, Chain-of-Thought prompting, and pretraining data composition.

Contribution

It provides a comprehensive assessment of state-of-the-art MLLMs, demonstrating how few-shot learning and prompting strategies enhance their reasoning and understanding abilities.

Findings

ICL and CoT prompting improve model performance

Pretraining on captioning datasets enhances zero-shot performance

Interleaved image-text pretraining benefits few-shot learning

Abstract

The linguistic capabilities of Multimodal Large Language Models (MLLMs) are critical for their effective application across diverse tasks. This study aims to evaluate the performance of MLLMs on the VALSE benchmark, focusing on the efficacy of few-shot In-Context Learning (ICL), and Chain-of-Thought (CoT) prompting. We conducted a comprehensive assessment of state-of-the-art MLLMs, varying in model size and pretraining datasets. The experimental results reveal that ICL and CoT prompting significantly boost model performance, particularly in tasks requiring complex reasoning and contextual understanding. Models pretrained on captioning datasets show superior zero-shot performance, while those trained on interleaved image-text data benefit from few-shot learning. Our findings provide valuable insights into optimizing MLLMs for better grounding of language in visual contexts, highlighting the importance of the composition of pretraining data and the potential of few-shot learning strategies to improve the reasoning abilities of MLLMs.