Papersby GlyphMeta research papers
A company publishing mostly on multimodal and robotics, best known here for LLaMA.40 papers · 67,081 citations · latest Oct 2026
InteractiveWorld: A Verifiable Human-Centric Environment for Building Better Interactive MLLM AssistantsInteractiveWorld provides a verifiable simulation environment and benchmark to evaluate whether multimodal language models can reliably guide humans through physical tasks.
Meta46 upvotesRead →Native Action-Prior Learning from Videos for World Action ModelsNAVA-WAM learns robot action priors directly from observation-only videos, improving control performance without needing action-labeled data for pretraining.Meta88 upvotesRead →Sharpening Tax in Post-TrainingThis paper introduces a metric to quantify the loss in solution coverage after reinforcement learning and proposes a sampler to mitigate it.Meta1 citations111 upvotesRead →World Action Modeling with Progressive Visual PlanningProWAM improves robotic control by predicting sparse visual sub-goals to guide action generation, avoiding the need for dense video generation.Meta1 citations98 upvotesRead →Context Language ModelsContext Language Models manage their own context by treating it as an editable file, improving performance and efficiency.Meta5 citations47 upvotesRead →In-Flight KV Cache with Clean Anchors for Faster Autoregressive Video DiffusionFlashForward accelerates video diffusion by reusing in-flight KV caches and using sparse clean anchors to maintain structural consistency.Meta1 citations34 upvotesRead →WearableQA: A Benchmark for Health Reasoning over Real-World Wearable DataWearableQA is a benchmark of multiple-choice questions derived from real longitudinal wearable data that evaluates large language model reasoning across data and health dimensions.Meta43 upvotesRead →Towards Physics of Multimodal Pretraining: Knowledge Flow, Modality Synergy, Early Unification, and RecipesControlled experiments on when text, image understanding and image generation help or compete when trained together.Meta7 citations63 upvotesRead →HumanCLAW: Can Vision-Language Models Act Through a Body?HumanCLAW decouples high-level vision-language decisions from low-level motor execution to evaluate embodied action intelligence, revealing that current vision-language models lack embodied self-awareness.Meta4 citations75 upvotesRead →TUA-Bench: A Benchmark for General-Purpose Terminal-Use AgentsTUA-Bench presents a comprehensive benchmark for evaluating general-purpose terminal-use agents across diverse digital activities and specialized workflows, revealing significant performance gaps among current frontier agents.Meta9 citations48 upvotesRead →Think in Strokes, Not Pixels: Process-Driven Image Generation via Interleaved ReasoningProcess-driven image generation decomposes synthesis into iterative steps involving textual planning, visual drafting, textual reflection, and visual refinement, with step-wise supervision ensuring consistency and interpretability.Meta6 citations71 upvotesRead →V-JEPA 2.1: Unlocking Dense Features in Video Self-Supervised LearningV-JEPA 2.1 is a self-supervised model that learns dense visual representations for images and videos through a combination of dense predictive loss, deep self-supervision, multi-modal tokenizers, and effective scaling.Meta161 citations39 upvotesRead →ReMix: Reinforcement routing for mixtures of LoRAs in LLM finetuningResearchers address imbalance in routing weights of Mixture-of-LoRAs models by proposing Reinforcement Routing (ReMix), which uses non-learnable weights and reinforcement learning techniques to improve model expressiveness and performance.Meta3 citations26 upvotesRead →Beyond Language Modeling: An Exploration of Multimodal PretrainingControlled multimodal pretraining experiments reveal key insights about unified visual representations, data complementarity, world modeling emergence, and efficient scaling through mixture-of-experts architectures.Meta44 citations108 upvotesRead →AIRS-Bench: a Suite of Tasks for Frontier AI Research Science AgentsAIRS-Bench presents a comprehensive benchmark suite for evaluating LLM agents across diverse scientific domains, demonstrating current limitations while providing open-source resources for advancement.Meta30 citations83 upvotesRead →Teaching Models to Teach Themselves: Reasoning at the Edge of LearnabilityA self-improvement framework enables pretrained language models to generate automated curricula for solving previously unsolvable problems by leveraging latent knowledge and meta-reinforcement learning.Meta25 citations43 upvotesRead →ShapeR: Robust Conditional 3D Shape Generation from Casual CapturesShapeR generates high-fidelity 3D shapes from casual image sequences using visual-inertial SLAM, 3D detection, and vision-language models with rectified flow transformer conditioning.Meta27 citations25 upvotesRead →Action100M: A Large-scale Video Action DatasetAction100M is a large-scale video action dataset constructed from internet instructional videos using automated pipelines with V-JEPA embeddings and GPT-based reasoning for structured annotations.Meta16 citations33 upvotesRead →VideoAuto-R1: Video Auto Reasoning via Thinking Once, Answering TwiceVideoAuto-R1 framework employs a reason-when-necessary strategy for video understanding, using a Thinking Once, Answering Twice training paradigm with verifiable rewards and confidence-based reasoning activation during inference.Meta23 citations38 upvotesRead →Physics of Language Models: Part 4.1, Architecture Design and the Magic of Canon LayersControlled synthetic pretraining reveals CANON LAYERS, lightweight architectural components that enhance information flow and model capabilities across various sequence architectures.Meta48 citations29 upvotesRead →SAM Audio: Segment Anything in AudioSAM Audio, a diffusion transformer-based foundation model, achieves superior performance in general audio separation using unified text, visual, and temporal span prompts across various audio types.Meta57 citations25 upvotesRead →OneStory: Coherent Multi-Shot Video Generation with Adaptive MemoryOneStory generates coherent multi-shot videos by modeling global cross-shot context through a Frame Selection module and an Adaptive Conditioner, leveraging pretrained image-to-video models and a curated dataset.Meta40 citations47 upvotesRead →Scaling Zero-Shot Reference-to-Video GenerationSaber is a scalable zero-shot framework for reference-to-video generation that uses video-text pairs to learn identity-consistent representations and outperforms models trained with explicit reference data.Meta11 citations29 upvotesRead →TUNA: Taming Unified Visual Representations for Native Unified Multimodal ModelsTUNA, a unified multimodal model, uses a cascaded VAE and representation encoder for end-to-end multimodal understanding and generation, outperforming decoupled models and achieving state-of-the-art results across various benchmarks.Meta38 citations78 upvotesRead →SAM 3: Segment Anything with ConceptsSegment Anything Model 3 achieves state-of-the-art performance in promptable concept segmentation and tracking by leveraging a unified model architecture with decoupled recognition and localization.Meta1,466 citations138 upvotesRead →SAM 3D: 3Dfy Anything in ImagesSAM 3D is a generative model that reconstructs 3D objects from single images using a multi-stage training framework that includes synthetic pretraining and real-world alignment, achieving high performance in human preference tests.Meta343 citations117 upvotesRead →What Does It Take to Be a Good AI Research Agent? Studying the Role of Ideation DiversityIdeation diversity significantly enhances the performance of AI research agents across various models and scaffolds on the MLE-bench benchmark.Meta8 citations57 upvotesRead →Souper-Model: How Simple Arithmetic Unlocks State-of-the-Art LLM PerformanceSoCE is a model souping approach that uses benchmark composition to identify and combine expert models for improved performance.Meta4 citations140 upvotesRead →The Path Not Taken: RLVR Provably Learns Off the PrincipalsReinforcement Learning with Verifiable Rewards (RLVR) improves large language models by updating a limited set of parameters, which is explained by a Three-Gate Theory, revealing distinct optimization dynamics compared to supervised fine-tuning.Meta65 citations37 upvotesRead →Scaling Agent Learning via Experience SynthesisDreamGym is a unified framework that synthesizes diverse experiences for scalable online RL training, improving agent performance and reducing real-world interactions.Meta52 citations83 upvotesRead →The Art of Scaling Reinforcement Learning Compute for LLMsA systematic study defines a framework for analyzing and predicting reinforcement learning scaling in large language models, identifying key design choices that affect compute efficiency and proposing a best-practice recipe.Meta112 citations34 upvotesRead →Agent Learning via Early ExperienceEarly experience, using agent-generated interaction data without reward signals, improves policy effectiveness and generalization, serving as a bridge between imitation learning and reinforcement learning.Meta79 citations256 upvotesRead →The Alignment Waltz: Jointly Training Agents to Collaborate for SafetyWaltzRL, a multi-agent reinforcement learning framework, improves LLM safety and helpfulness by collaboratively training a conversation agent and a feedback agent, reducing unsafe responses and overrefusals.Meta13 citations41 upvotesRead →TruthRL: Incentivizing Truthful LLMs via Reinforcement LearningTruthRL, a reinforcement learning framework, enhances the truthfulness of large language models by balancing accuracy and abstention, significantly reducing hallucinations and improving performance across benchmarks.Meta24 citations56 upvotesRead →Jointly Reinforcing Diversity and Quality in Language Model GenerationsDARLING, a diversity-aware reinforcement learning framework, enhances both the quality and diversity of large language model outputs across various tasks.Meta92 citations25 upvotesRead →DINOv3DINOv3, a self-supervised learning model, achieves superior performance across various vision tasks by scaling datasets and models, addressing dense feature degradation, and enhancing flexibility with post-hoc strategies.Meta1,932 citations320 upvotesRead →VGGT: Visual Geometry Grounded TransformerVGGT, a feed-forward neural network, efficiently infers multiple 3D attributes from single or multiple views, outperforming alternatives and enhancing downstream tasks without post-processing.Meta2,132 citations40 upvotesRead →The Llama 3 Herd of ModelsOpen models up to 405B parameters that match leading closed models on many tasks.Meta20.8k citations119 upvotesRead →Segment AnythingA promptable model and a dataset of over a billion masks that cut out any object in any image.Meta16.8k citations6 upvotesRead →LLaMA: Open and Efficient Foundation Language ModelsOpen models from 7B to 65B trained only on public data; the 13B model beats GPT-3 on most benchmarks.Meta22.6k citations25 upvotesRead →
About and links
Topics: Multimodal and robotics 15, Agents and evaluation 5, Reasoning 4, Architectures 3, Training and scaling 3, Foundation models 3, Alignment and safety 2, Inference and efficiency 2, Retrieval and data 2, Applied AI 1
Meta lab profile on Labs (labs.fru.dev): leaders, models, landmarksMeta on Companies (companies.fru.dev)Also onAcquisitions·Quarterly·Paydays·TechConf