We present Magma, a foundation model that serves multimodal AI agentic tasks in both the digital and physical worlds. Magma is a significant extension of vision-language (VL) models in that it not only retains the VL understanding ability (verbal intelligence) of the latter, but is also equipped with the ability to ground and act in the visual-spatial world (spatial-temporal intelligence). To endow agentic capabilities for tasks ranging from UI navigation to robot manipulation, Magma is trained on large amounts of heterogeneous datasets that span from images, videos to robotics data, where actionable visual objects (e.g. clickable buttons in GUI) in images are labeled by Set-of-Mark (SoM) for action grounding, and object movements (e.g. trace of human hands or robotic arms) in videos are labeled by Trace-of-Mark (ToM) for action planning. Extensive experiments show that SoM and ToM help bridge the gap between verbal and action abilities and significantly enhance spatio-temporal intelligence which is fundamental to agentic tasks, as shown in Fig. 1. In particular, Magma creates new state-of-the-art results on UI navigation and robotic manipulation tasks, outperforming previous models that are specifically tailored to these tasks. Moreover, Magma preserves strong multimodal understanding ability and compares favorably to popular large multimodal models that are trained on much larger datasets. We have made our model and code public for reproducibility1.

Large foundation models have shown strong open-world generalization to complex problems in vision and language, but similar levels of generalization have yet to be achieved in robotics. One fundamental challenge is the lack of robotic data, which are typically obtained through expensive on-robot operation. A promising remedy is to leverage cheaper, *off-domain* data such as action-free videos, hand-drawn sketches, or simulation data. In this work, we posit that *hierarchical* vision-language-action (VLA) models can be more effective in utilizing off-domain data than standard monolithic VLA models that directly finetune vision-language models (VLMs) to predict actions. In particular, we study a class of hierarchical VLA models, where the high-level VLM is finetuned to produce a coarse 2D path indicating the desired robot end-effector trajectory given an RGB image and a task description. The intermediate 2D path prediction is then served as guidance to the low-level, 3D-aware control policy capable of precise manipulation. Doing so alleviates the high-level VLM from fine-grained action prediction, while reducing the low-level policy's burden on complex task-level reasoning. We show that, with the hierarchical design, the high-level VLM can transfer across significant domain gaps between the off-domain finetuning data and real-robot testing scenarios, including differences in embodiments, dynamics, visual appearances, and task semantics, etc. In the real-robot experiments, we observe an average of 20% improvement in success rate across seven different axes of generalization over OpenVLA, representing a 50% relative gain. Visual results are provided at: [https://hamster-robot.github.io/](https://hamster-robot.github.io/)

In this work, we investigate how a model's tendency to broadly integrate its parametric knowledge evolves throughout pretraining, and how this behavior affects overall performance, particularly in terms of knowledge acquisition and forgetting. We introduce the concept of knowledge entropy, which quantifies the range of memory sources the model engages with; high knowledge entropy indicates that the model utilizes a wide range of memory sources, while low knowledge entropy suggests reliance on specific sources with greater certainty. Our analysis reveals a consistent decline in knowledge entropy as pretraining advances. We also find that the decline is closely associated with a reduction in the model's ability to acquire and retain knowledge, leading us to conclude that diminishing knowledge entropy (smaller number of active memory sources) impairs the model's knowledge acquisition and retention capabilities. We find further support for this by demonstrating that increasing the activity of inactive memory sources enhances the model's capacity for knowledge acquisition and retention.

We introduce Latent Action Pretraining for general Action models (LAPA), the first unsupervised method for pretraining Vision-Language-Action (VLA) models without ground-truth robot action labels. Existing Vision-Language-Action models require action labels typically collected by human teleoperators during pretraining, which significantly limits possible data sources and scale. In this work, we propose a method to learn from internet-scale videos that do not have robot action labels. We first train an action quantization model leveraging VQ-VAE-based objective to learn discrete latent actions between image frames, then pretrain a latent VLA model to predict these latent actions from observations and task descriptions, and finally finetune the VLA on small-scale robot manipulation data to map from latent to robot actions. Experimental results demonstrate that our method significantly outperforms existing techniques that train robot manipulation policies from large-scale videos. Furthermore, it outperforms the state-of-the-art VLA model trained with robotic action labels on real-world manipulation tasks that require language conditioning, generalization to unseen objects, and semantic generalization to unseen instructions. Training only on human manipulation videos also shows positive transfer, opening up the potential for leveraging web-scale data for robotics foundation models.

Seungone Kim, Juyoung Suk, Ji Yong Cho, Shayne Longpre, Chaeeun Kim, Dongkeun Yoon, Guijin Son, Yejin Cho, Sheikh Shafayat, Jinheon Baek, Sue Hyun Park, Hyeonbin Hwang, Jinkyung Jo, Hyowon Cho, Haebin Shin, Seongyun Lee, Hanseok Oh, Noah Lee, Namgyu Ho, Se June Joo, Miyoung Ko, Yoonjoo Lee, Hyungjoo Chae, Jamin Shin, Joel Jang, Seonghyeon Ye, Bill Yuchen Lin, Sean Welleck, Graham Neubig, Moontae Lee, Kyungjae Lee, Minjoon Seo. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Video understanding is fundamental to tasks such as action recognition, video reasoning, and robotic control. Early video understanding methods based on large vision-language models (LVLMs) typically adopt a single-pass reasoning paradigm without dynamic feedback, limiting the model’s capacity to self-correct and adapt in complex scenarios. Recent efforts have attempted to address this limitation by incorporating reward models and reinforcement learning to enhance reasoning, or by employing tool-agent frameworks. However, these approaches face several challenges, including high annotation costs, reward signals that fail to capture real-time reasoning states, and low inference efficiency. To overcome these issues, we propose ReAgent-V, a novel agentic video understanding framework that integrates efficient frame selection with real-time reward generation during inference. These reward signals not only guide iterative answer refinement through a multi-perspective reflection mechanism—adjusting predictions from conservative, neutral, and aggressive viewpoints—but also enable automatic filtering of high-quality data for supervised fine-tuning (SFT), direct preference optimization (DPO), and group relative policy optimization (GRPO). ReAgent-V is lightweight, modular, and extensible, supporting flexible tool integration tailored to diverse tasks. Extensive experiments on 12 datasets across three core applications—video understanding, video reasoning enhancement, and vision-language-action model alignment—demonstrate significant gains in generalization and reasoning, with improvements of up to 6.9%, 2.1%, and 9.8%, respectively, highlighting the effectiveness and versatility of the proposed framework.

Hyunji Lee, Doyoung Kim, Jihoon Jun, Se June Joo, Joel Jang, Kyoung-Woon On, Minjoon Seo. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Dongkeun Yoon, Joel Jang, Sungdong Kim, Seungone Kim, Sheikh Shafayat, Minjoon Seo. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Benchmarking the performance of information retrieval (IR) is mostly conducted with a fixed set of documents (static corpora).However, in realistic scenarios, this is rarely the case and the documents to be retrieved are constantly updated and added.In this paper, we focus on Generative Retrievals (GR), which apply autoregressive language models to IR problems, and explore their adaptability and robustness in dynamic scenarios.We also conduct an extensive evaluation of computational and memory efficiency, crucial factors for real-world deployment of IR systems handling vast and ever-changing document collections.Our results on the StreamingQA benchmark demonstrate that GR is more adaptable to evolving knowledge (4 -11%), robust in learning knowledge with temporal information, and efficient in terms of inference FLOPs (ˆ2), indexing time (ˆ6), and storage footprint (ˆ4) compared to Dual Encoders (DE), which are commonly used in retrieval systems.Our paper highlights the potential of GR for future use in practical IR systems within dynamic environments.

Recently, GPT-4 has become the de facto evaluator for long-form text generated by large language models (LLMs). However, for practitioners and researchers with large and custom evaluation tasks, GPT-4 is unreliable due to its closed-source nature, uncontrolled versioning, and prohibitive costs. In this work, we propose PROMETHEUS a fully open-source LLM that is on par with GPT-4’s evaluation capabilities when the appropriate reference materials (reference answer, score rubric) are accompanied. For this purpose, we construct a new dataset – FEEDBACK COLLECTION – that consists of 1K fine-grained score rubrics, 20K instructions, and 100K natural language feedback generated by GPT-4. Using the FEEDBACK COLLECTION, we train PROMETHEUS, a 13B evaluation-specific LLM that can assess any given response based on novel and unseen score rubrics and reference materials provided by the user. Our dataset’s versatility and diversity make our model generalize to challenging real-world criteria, such as prioritizing conciseness, child-readability, or varying levels of formality. We show that PROMETHEUS shows a stronger correlation with GPT-4 evaluation compared to ChatGPT on seven evaluation benchmarks (Two Feedback Collection testsets, MT Bench, Vicuna Bench, Flask Eval, MT Bench Human Judgment, and HHH Alignment), showing the efficacy of our model and dataset design. During human evaluation with hand-crafted score rubrics, PROMETHEUS shows a Pearson correlation of 0.897 with human evaluators, which is on par with GPT-4-0613 (0.882), and greatly outperforms ChatGPT (0.392). Remarkably, when assessing the quality of the generated feedback, PROMETHEUS demonstrates a win rate of 58.62% when compared to GPT-4 evaluation and a win rate of 79.57% when compared to ChatGPT evaluation. Our findings suggests that by adding reference materials and training on GPT-4 feedback, we can obtain effective open-source evaluator LMs.

Hyunji Lee, Se June Joo, Chaeeun Kim, Joel Jang, Doyoung Kim, Kyoung-Woon On, Minjoon Seo. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.

Abstract Previous work in prompt engineering for large language models has introduced different gradient-free probability-based prompt selection methods that aim to choose the optimal prompt among the candidates for a given task but have failed to provide a comprehensive and fair comparison between each other. In this paper, we propose a unified framework to interpret and evaluate the existing probability-based prompt selection methods by performing extensive experiments on 13 common and diverse NLP tasks. We find that each of the existing methods can be interpreted as some variant of the method that maximizes mutual information between the input and the predicted output (MI). Utilizing this finding, we develop several other combinatorial variants of MI and increase the effectiveness of the oracle prompt selection method from 87.79% to 94.98%, measured as the ratio of the performance of the selected prompt to that of the optimal oracle prompt. Furthermore, considering that all the methods rely on the output probability distribution of the model that might be biased, we propose a novel calibration method called Calibration by Marginalization (CBM) that is orthogonal to the existing methods and helps increase the prompt selection effectiveness of the best method to 96.85%, achieving 99.44% of the oracle prompt F1 without calibration.1

Joel Jang, Dongkeun Yoon, Sohee Yang, Sungmin Cha, Moontae Lee, Lajanugen Logeswaran, Minjoon Seo. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

Language models (LMs) with less than 100B parameters are known to perform poorly on chain-of-thought (CoT) reasoning in contrast to large LMs when solving unseen tasks.In this work, we aim to equip smaller LMs with the step-by-step reasoning capability by instruction tuning with CoT rationales.In order to achieve this goal, we first introduce a new instructiontuning dataset called the COT COLLECTION, which augments the existing Flan Collection (including only 9 CoT tasks) with additional 1.84 million rationales across 1,060 tasks.We show that CoT fine-tuning Flan-T5 (3B & 11B) with COT COLLECTION enables smaller LMs to have better CoT capabilities on unseen tasks.On the BIG-Bench-Hard (BBH) benchmark, we report an average improvement of +4.34% (Flan-T5 3B) and +2.60% (Flan-T5 11B), in terms of zero-shot task accuracy.Furthermore, we show that instruction tuning with COT COL-LECTION allows LMs to possess stronger fewshot learning capabilities on 4 domain-specific tasks, resulting in an improvement of +2.24% (Flan-T5 3B) and +2.37% (Flan-T5 11B), even outperforming ChatGPT utilizing demonstrations until the max length by a +13.98% margin.Our code, the COT COLLECTION data, and model checkpoints are publicly available 1 .

Recently, Language Models (LMs) instruction-tuned on multiple tasks, also known as multitask-prompted fine-tuning (MT), have shown capabilities to generalize to unseen tasks. Previous work has shown that scaling the number of finetuning datasets and instructions is the key component in making stronger MT LMs. In this work, we report surprising findings that show an expert LM trained on just a single task can outperform an MT LM trained with 300+ different tasks on 11 different unseen datasets and on 13 datasets of the BIG-bench benchmark by an average of 3.20% and 1.29%, respectively. This finding casts doubt on the previously held belief that simply scaling the number of tasks makes stronger MT LMs. Leveraging this finding, we further show that this distributed approach of training multiple expert LMs instead of a single MT LM for zero-shot inference possesses many benefits including (1) avoiding negative task transfer that often occurs during instruction tuning, (2) being able to continually learn new tasks without having to re-train on previous tasks to avoid catastrophic forgetting, and (3) showing compositional capabilities when merging individual experts together.

Language Models (LMs) become outdated as the world changes; they often fail to perform tasks requiring recent factual information which was absent or different during training, a phenomenon called temporal misalignment.This is especially a challenging problem because the research community still lacks a coherent dataset for assessing the adaptability of LMs to frequently-updated knowledge corpus such as Wikipedia.To this end, we introduce TEMPORALWIKI, a lifelong benchmark for ever-evolving LMs that utilizes the difference between consecutive snapshots of English Wikipedia and English Wikidata for training and evaluation, respectively.The benchmark hence allows researchers to periodically track an LM's ability to retain previous knowledge and acquire updated/new knowledge at each point in time.We also find that training an LM on the diff data through continual learning methods achieves similar or better perplexity than on the entire snapshot in our benchmark with 12 times less computational cost, which verifies that factual knowledge in LMs can be safely updated with minimal training data via continual learning.The dataset and the code is made available at this link. What is the most dominant COVID-19 variant?What is the most dominant COVID-19 variant?Difference