每天自动聚合 AI 领域最新动态
Reusing the attention KV cache across requests is a standard way to cut prefill cost in long-context LLM serving, and...
We introduce Gen4P, a unified framework for recovering 4D geometry and motion from monocular video. We show that a pr...
Reordering multiple-choice options can change a model's answer, an effect studied as selection bias (a preference for...
Reinforcement learning with verifiable rewards (RLVR) typically uses small rollout groups per prompt, where repeated ...
Vision-language-action (VLA) models map visual observations and language instructions to continuous robot actions. Th...
Personalization of Large Language Models (LLMs) is increasingly important for user-centric applications, yet prior me...
Transformers are intuitively considered to be efficient at copying from their inputs. We actually show that the abili...
When does a language model acquire a capability? Developmental studies usually answer this by tracking when models st...
Time series foundation models are becoming increasingly popular for general-purpose forecasting. However, most leadin...
Character Error Rate (CER), the standard metric for OCR and vision-language models (VLMs), charges a typo as much as ...
AI systems now write long and difficult mathematical proofs, and checking their step-by-step validity is becoming a b...
Agent environments are essential for large language model (LLM) development. They use programmatic verifiers to turn ...