TLDR: Token-Level Detective Reward Model for Large Vision Language Models

TL;DR

This paper introduces TLDR, a token-level reward model for large vision-language models that provides fine-grained feedback, improves model performance, and accelerates human annotation processes.

Contribution

The paper proposes a novel token-level reward model with a perturbation-based training method, enhancing interpretability and utility in vision-language tasks.

Findings

TLDR improves model performance significantly.

It assists in self-correction of generation outputs.

It speeds up human annotation by 3 times.

Abstract

Although reward models have been successful in improving multimodal large language models, the reward models themselves remain brutal and contain minimal information. Notably, existing reward models only mimic human annotations by assigning only one binary feedback to any text, no matter how long the text is. In the realm of multimodal language models, where models are required to process both images and texts, a naive reward model may learn implicit biases toward texts and become less grounded in images. In this paper, we propose a $\textbf{T}$oken-$\textbf{L}$evel $\textbf{D}$etective $\textbf{R}$eward Model ($\textbf{TLDR}$) to provide fine-grained annotations to each text token. We first introduce a perturbation-based method to generate synthetic hard negatives and their token-level labels to train TLDR models. Then we show the rich usefulness of TLDR models both in assisting off-the-shelf models to self-correct their generations, and in serving as a hallucination evaluation tool. We show that TLDR automatically trains a token-level likelihood optimization, and can improve the base model's performance significantly. Finally, we show that TLDR models can significantly speed up human annotation by 3 times to acquire a broader range of high-quality vision language data.