REVERSE-LLaVA-MORE-8B is an open-source vision-language model (VLM) that performs both next-token prediction and self-verification / self-correction during generation. It is built upon LLaVA-MORE (LLaVA with LLaMA-3.1) and fine-tuned on the REVERSE Visual Instruct 1.3M dataset. The model is equipped with a retrospective resampling mechanism that detects and corrects hallucinations on the fly. Training was conducted in early March, 2025.

REVERSE-LLaVA-MORE-8B delivers strong performance gains in hallucination reduction across multiple captioning and open-ended VQA benchmarks:

On discriminative tasks, REVERSE-LLaVA-MORE performs competitively with existing base VLM:

Please refer to the installation guide on GitHub to get started:
👉 Installation Guide

Target Users:
Researchers, developers, and students working on VLMs, hallucination mitigation, and vision-language alignment.