Visual Instruction Tuning with Polite Flamingo
This addresses the problem of reduced human preference in multi-modal LLMs for vision-language tasks, though it appears incremental as it builds on existing fine-tuning approaches.
The paper tackles the 'multi-modal alignment tax' problem where vision-language models lose politeness during fine-tuning due to raw annotations, by introducing Polite Flamingo to rewrite responses into polite formats and creating the PF-1M dataset. The resulting Clever Flamingo model shows improved multi-modal understanding and response politeness in evaluations.
Recent research has demonstrated that the multi-task fine-tuning of multi-modal Large Language Models (LLMs) using an assortment of annotated downstream vision-language datasets significantly enhances their performance. Yet, during this process, a side effect, which we termed as the "multi-modal alignment tax", surfaces. This side effect negatively impacts the model's ability to format responses appropriately -- for instance, its "politeness" -- due to the overly succinct and unformatted nature of raw annotations, resulting in reduced human preference. In this paper, we introduce Polite Flamingo, a multi-modal response rewriter that transforms raw annotations into a more appealing, "polite" format. Polite Flamingo is trained to reconstruct high-quality responses from their automatically distorted counterparts and is subsequently applied to a vast array of vision-language datasets for response rewriting. After rigorous filtering, we generate the PF-1M dataset and further validate its value by fine-tuning a multi-modal LLM with it. Combined with novel methodologies including U-shaped multi-stage tuning and multi-turn augmentation, the resulting model, Clever Flamingo, demonstrates its advantages in both multi-modal understanding and response politeness according to automated and human evaluations.