HuggingFace
Compute Globally, Materialize Locally: The Memory Contract of Sparse Event-KV
Long-horizon agents increasingly reuse their KV cache as memory: a serving system keeps a subset of cached entries an...
🤖 AI 前沿资讯
每天自动聚合 AI 领域最新动态
HuggingFace
Seeing or Knowing? Visual Context Sensitivity in Multimodal Large Language Models
Multimodal Large Language Models (MLLMs) achieve strong performance by integrating visual inputs with the rich priors...
HuggingFace
Loud or Silent? A Reusable Framework for Per-Modality Failure Analysis in Multimodal Clinical AI
Multimodal clinical models are usually judged on accuracy with every modality present, but deployment removes modalit...
HuggingFace
Zero-Mem: Zero-Token Memory Operations for LLM Agents
LLM agents need memory to act consistently over long interactions, yet many systems use additional LLM calls to opera...
HuggingFace
SG-WAM: Self-Guided World Modeling in Geometry-Aware Policy Space
World Action Models (WAMs) couple action generation with prediction of future states. Their effectiveness depends on ...
HuggingFace
To Add Is Machine, To Delete Is Human: Measuring and Mitigating Deletion Avoidance in LLM Code Editing
Large language models increasingly write and repair production code, yet evidence is mounting that their test-passing...
HuggingFace
Wnuan: Staged Post-Training for Question Answering over Proprietary Enterprise Knowledge
Enterprise question answering requires models to acquire proprietary knowledge without discarding general capabilitie...
HuggingFace
MemSFT: Mitigating Alignment Tax with an External Parametric Memory
Adapting Large Language Models (LLMs) to specialized domains often incurs an alignment tax, as fine-tuning on domain-...
HuggingFace
AgentStream: How Well Do Self-Evolving LLM Agents Perform Under Streaming Tasks?
Large language model (LLM) agents can self-evolve by continually improving from their own accumulated experience. How...
arXiv
Long-term Measurements: Towards a Longitudinal Understanding of Human-AI Interactions
Language models have taken on the role of a very new type of technology, by virtue of their "human-ness" and rapid in...
arXiv
DyFrDet: Towards Accurate Small Object Detection via Dynamic Frequency Suppression with Label Disambiguation
Despite the remarkable progress over the past decades, accurately identifying small objects remains challenging becau...
arXiv
SWE-Touch: Benchmarking Coding Agents When Users Touch the Code
Real-world software development requires coding agents to operate in shared workspaces where users may inspect and mo...