--- title: README emoji: ⚡ colorFrom: blue colorTo: red sdk: static pinned: false --- - **[2024-11]** 🤯🤯 We introduce **Multimodal SAE**, the first framework designed to interpret learned features in large-scale multimodal models using Sparse Autoencoders. Through our approach, we leverage LLaVA-OneVision-72B to analyze and explain the SAE-derived features of LLaVA-NeXT-LLaMA3-8B. Furthermore, we demonstrate the ability to steer model behavior by clamping specific features to alleviate hallucinations and avoid safety-related issues. [GitHub](https://github.com/EvolvingLMMs-Lab/multimodal-sae) - **[2024-10]** 🔥🔥 We present **`LLaVA-Critic`**, the first open-source large multimodal model as a generalist evaluator for assessing LMM-generated responses across diverse multimodal tasks and scenarios. [GitHub](https://github.com/LLaVA-VL/LLaVA-NeXT) | [Blog](https://llava-vl.github.io/blog/2024-10-03-llava-critic/) - **[2024-10]** 🎬🎬 Introducing **`LLaVA-Video`**, a family of open large multimodal models designed specifically for advanced video understanding. We're open-sourcing **LLaVA-Video-178K**, a high-quality, synthetic dataset for video instruction tuning. [GitHub](https://github.com/LLaVA-VL/LLaVA-NeXT) | [Blog](https://github.com/LLaVA-VL/LLaVA-NeXT) - **[2024-08]** 🤞🤞 We present **`LLaVA-OneVision`**, a family of LMMs developed by consolidating insights into data, models, and visual representations. [GitHub](https://github.com/LLaVA-VL/LLaVA-NeXT) | [Blog](https://llava-vl.github.io/blog/2024-08-05-llava-onevision/) - **[2024-06]** 🧑‍🎨🧑‍🎨 We release **`LLaVA-NeXT-Interleave`**, an LMM extending capabilities to real-world settings: Multi-image, Multi-frame (videos), Multi-view (3D), and Multi-patch (single-image). [GitHub](https://github.com/LLaVA-VL/LLaVA-NeXT) | [Blog](https://llava-vl.github.io/blog/2024-06-16-llava-next-interleave/) - **[2024-06]** 🚀🚀 We release **`LongVA`**, a long language model with state-of-the-art video understanding performance. [GitHub](https://github.com/EvolvingLMMs-Lab/LongVA) | [Blog](https://lmms-lab.github.io/posts/longva/)
Older Updates (2024-06 and earlier) - **[2024-06]** 🎬🎬 The **`lmms-eval/v0.2`** toolkit now supports video evaluations for models like LLaVA-NeXT Video and Gemini 1.5 Pro. [GitHub](https://github.com/EvolvingLMMs-Lab/lmms-eval) | [Blog](https://lmms-lab.github.io/posts/lmms-eval-0.2/) - **[2024-05]** 🚀🚀 We release **`LLaVA-NeXT Video`**, a model performing at Google's Gemini level on video understanding tasks. [GitHub](https://github.com/LLaVA-VL/LLaVA-NeXT) | [Blog](https://llava-vl.github.io/blog/2024-04-30-llava-next-video/) - **[2024-05]** 🚀🚀 The **`LLaVA-NeXT`** model family reaches near GPT-4V performance on multimodal benchmarks, with models up to 110B parameters. [GitHub](https://github.com/LLaVA-VL/LLaVA-NeXT) | [Blog](https://llava-vl.github.io/blog/2024-05-10-llava-next-stronger-llms/) - **[2024-03]** We release **`lmms-eval`**, a toolkit for holistic evaluations with 50+ multimodal datasets and 10+ models. [GitHub](https://github.com/EvolvingLMMs-Lab/lmms-eval) | [Blog](https://lmms-lab.github.io/posts/lmms-eval-0.1/)