Hugging Face · HN · arXiv

Tendencias

Lo que se publica y lo que se discute: Daily Papers de Hugging Face y Hacker News.

Papers del día

  • BDH-CQ: In-Context Learning with Recurrent Latent Reasoning2608.09888We introduce BDH-CQ, a reasoning model that combines in-context learning with recurrent latent reasoning. Inputs presented at inference time continuously update the model's recurrent memory; the model then solves a query through iterative computation in a high-dimensional latent space, without verbalizing its intermediate reasoning. We evaluate the model on…
  • LLM-as-a-Verifier: A General-Purpose Verification Framework2607.05391Scaling pre-training, post-training, and test-time compute have become the central paradigms for improving the capabilities of LLMs. In this work, we identify verification, the ability to determine the correctness of a solution, as a new scaling axis. To unlock this and demonstrate its effectiveness, we introduce LLM-as-a-Verifier, a general-purpose verific…
  • COLLEAGUE.SKILL: Automated AI Skill Generation via Expert Knowledge Distillation2605.31264LLM agents are increasingly expected not only to complete isolated tasks, but also to carry bounded representations of human expertise, judgment, and interaction style. Building such person-grounded agents remains difficult because actionable knowledge associated with a person or role is usually embedded in heterogeneous traces rather than written as clean …
  • Securing the AI Agent: A Unified Framework for Multi-Layer Agent Red Teaming2606.31227The fast growth of open-source AI infrastructure, from model serving engines and agent platforms to the Model Context Protocol (MCP) ecosystem and the language models themselves, has outpaced the security tooling available to defend it. We present AI-Infra-Guard, an open-source framework that organizes AI red teaming around a single observation: the attack …
  • TradingAgents: Multi-Agents LLM Financial Trading Framework2412.20138Significant progress has been made in automated problem-solving using societies of agents powered by large language models (LLMs). In finance, efforts have largely focused on single-agent systems handling specific tasks or multi-agent frameworks independently gathering data. However, the multi-agent systems' potential to replicate real-world trading firms' …
  • HarnessEval-W: Agentifying the Evaluation of Visual Worlds2608.16859A benchmark should deliver more than a scalar score: what makes an evaluation trustworthy is the reasoning that justifies the score. This is especially critical for world models, where judging a rollout requires understanding whether physics, causality, and world state evolve correctly. Humans spot such violations naturally, yet no existing benchmark automa…
  • OpenDevin: An Open Platform for AI Software Developers as Generalist Agents2407.16741Software is one of the most powerful tools that we humans have at our disposal; it allows a skilled programmer to interact with the world in complex and profound ways. At the same time, thanks to improvements in large language models (LLMs), there has also been a rapid development in AI agents that interact with and affect change in their surrounding enviro…
  • Unlimited OCR Works2606.23050Recently, end-to-end OCR models, exemplified by DeepSeek OCR, have once again thrust OCR into the spotlight. A widely held view is that employing a large language model (LLM) as the decoder allows the model to leverage the prior distribution of language, leading to improved OCR performance. However, the downside is equally evident: as the output sequence le…
  • SmolDocling: An ultra-compact vision-language model for end-to-end multi-modal document conversion2503.11576We introduce SmolDocling, an ultra-compact vision-language model targeting end-to-end document conversion. Our model comprehensively processes entire pages by generating DocTags, a new universal markup format that captures all page elements in their full context with location. Unlike existing approaches that rely on large foundational models, or ensemble so…
  • Efficient Memory Management for Large Language Model Serving with PagedAttention2309.06180High throughput serving of large language models (LLMs) requires batching sufficiently many requests at a time. However, existing systems struggle because the key-value cache (KV cache) memory for each request is huge and grows and shrinks dynamically. When managed inefficiently, this memory can be significantly wasted by fragmentation and redundant duplica…
  • MinerU2.5: A Decoupled Vision-Language Model for Efficient High-Resolution Document Parsing2509.22186We introduce MinerU2.5, a 1.2B-parameter document parsing vision-language model that achieves state-of-the-art recognition accuracy while maintaining exceptional computational efficiency. Our approach employs a coarse-to-fine, two-stage parsing strategy that decouples global layout analysis from local content recognition. In the first stage, the model perfo…
  • Kronos: A Foundation Model for the Language of Financial Markets2508.02739The success of large-scale pre-training paradigm, exemplified by Large Language Models (LLMs), has inspired the development of Time Series Foundation Models (TSFMs). However, their application to financial candlestick (K-line) data remains limited, often underperforming non-pre-trained architectures. Moreover, existing TSFMs often overlook crucial downstrea…
  • Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory2504.19413Large Language Models (LLMs) have demonstrated remarkable prowess in generating contextually coherent responses, yet their fixed context windows pose fundamental challenges for maintaining consistency over prolonged multi-session dialogues. We introduce Mem0, a scalable memory-centric architecture that addresses this issue by dynamically extracting, consoli…
  • MatrAIx: Simulating the World with 8.3 Billion Persona Agents2608.04205Human evaluation of AI systems and digital products is costly, slow, and difficult to scale. Offline evaluations are more scalable but often abstract away human diversity and interactive behavior. We therefore introduce MatrAIx, a population-scale simulated-user evaluation infrastructure for testing AI systems and digital products with heterogeneous users. …
  • StateM: Reaching 95.3% Raw Accuracy, or a \$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling2608.15089Long-horizon agents can fail even when their underlying models can solve the constituent steps. They may lose track of mutable state, fail to reactivate lessons from earlier executions, skip known procedures, or stop prematurely. We bet on harness scaling to improve the execution system around an agent without changing its model weights. We introduce StateM…
  • Zetta ζ: An Efficient Closed-Loop Embodied Harness for Self-Evolving Physical Intelligence2608.16590Embodied agents are increasingly used to close the gap left by end-to-end policy models. Yet the agentic path has not realized closed-loop learning in physical execution: existing harnesses remain largely open-loop, following fixed skills during rollout and reflecting only after an episode completes. Such post-hoc reflection cannot govern execution as it un…

Hacker News

Open-weight con más likes

  • deepseek-ai/DeepSeek-R113,582
  • meta-llama/Llama-3.1-8B-Instruct6,641
  • meta-llama/Meta-Llama-3-8B6,630
  • deepseek-ai/DeepSeek-V4-Pro5,464
  • openai/gpt-oss-120b5,114
  • bigscience/bloom5,038
  • zai-org/GLM-5.25,014
  • openai/gpt-oss-20b4,935

Open-weight en Hugging Face

Modelolikesdownloads
deepseek-ai/DeepSeek-R113,5826,388,560
meta-llama/Llama-3.1-8B-Instruct6,6416,888,821
meta-llama/Meta-Llama-3-8B6,6301,504,603
deepseek-ai/DeepSeek-V4-Pro5,4641,182,759
openai/gpt-oss-120b5,1144,680,489
bigscience/bloom5,03817,082
zai-org/GLM-5.25,0142,765,508
openai/gpt-oss-20b4,9357,590,370
meta-llama/Meta-Llama-3-8B-Instruct4,8291,849,725
meta-llama/Llama-2-7b-chat-hf4,822787,485
meta-llama/Llama-2-7b4,523116
deepseek-ai/DeepSeek-V34,1741,039,428
mistralai/Mistral-7B-v0.14,144472,475
deepseek-ai/DeepSeek-V4-Flash-07313,5832,547,549
microsoft/phi-23,5011,624,322
openai-community/gpt23,40914,188,958

Snapshot de build: 21 ago 2026