🚨 New paper alert! We introduce InterleavedBench📚, the first comprehensive evaluation benchmark for interleaved text-and-image generation, as well as InterleavedEval🔍, a powerful GPT-based evaluator that supports multi-aspect assessment. arXiv: (1/n)

Teach GPT-4o to edit on charts and tables to ReFocus 🔍 and facilitate reasoning 🧠! 🔥 We introduce ReFocus, which edits input table and chart images to better reason visually 🤔 Can we teach smaller models to learn such visual CoT reasoning? 🚀 Yes -- They are better than QA and CoT data! 📈 ReFocus + GPT-4o brings +11.0% on tables and +6.8% on charts without using any tools🔧! 📊 We release a 14K Visual CoT Reasoning *Training Dataset* that provides intermediate refocusing supervision. 🤖+🔍 > CoT: ReFocus VCoT is 8.0% better than QA data and 2.6% better CoT data with supervised Finetuning on Phi3.5v. Trained model also released. 📑 Check out This work is done during intern @Microsoft with amazing coauthors @minqian_liu @zhengyuan_yang @JCorring36990 @YijuanLu @jw2yang4ai @DanRothNLP @DineiFlorencio @ChaZhang. A huge shoutout to everyone!

Can Text-to-Image models understand common sense? 🤔 Can they generate images that fit everyday common sense? 🤔 tldr; NO, they are far less intelligent than us 💁🏻‍♀️ Introducing Commonsense-T2I 💡 a novel evaluation and benchmark designed to measure commonsense reasoning in T2I models 🔥🔥 Paper: (1/n)