HuggingFace
Reinforcing Agentic Creativity in Scientific Ideation with Night Science
Large language models (LLMs) excel at structured, verifiable tasks, but their low-entropy bias can produce homogeneou...
🤖 AI 前沿资讯
每天自动聚合 AI 领域最新动态
HuggingFace
VisionHOPE: Visual Backbones as Self-Modifying Learning Systems
Visual backbones have evolved from Convolutional Neural Networks (CNNs) with local aggregation to Vision Transformers...
HuggingFace
Who Gets a Token, and What Does It Carry? Unequal Name Support and Concept Access in Large Language Models
Names are personal identifiers, but they also carry social meaning and are widely used to evaluate how language model...
HuggingFace
SentZero: An Enhanced Sentence-Centric Vision-Language Pretraining for Multi-Task Zero-Shot Chest X-Ray Analysis
Vision-language (VL) pretraining using paired chest X-ray (CXR) images and radiology reports has shown strong potenti...
HuggingFace
Allspark: Weak to Strong Transfer via Alternating Chain of Thought
Recent progress in frontier models has renewed interest in large-scale reinforcement learning (RL), but the cost of g...
HuggingFace
ExpVoyager: Direct Experience Navigation for Dynamic Agent Skill Synthesis
Learning from experience in LLM agents has become a key paradigm for developing self-evolving agents that continuousl...
HuggingFace
Can We Trust the Teacher? Decoupled Credit Direction-Magnitude for Self-Distillation
RLVR provides reliable trajectory-level credit, while OPSD offers dense supervision for token-level credit. This expo...
HuggingFace
WhiteMatter: All-to-All Cross-Layer Connections via KV Source Mixing
When generating text, a Transformer produces representations of past tokens at every layer, but each layer can normal...
HuggingFace
Do We Really Need KL Divergence for On-Policy Distillation of Large Language Models?
Since the advent of knowledge distillation, KL divergence has been the standard loss in distillation. Recently, on-po...
HuggingFace
NVAlign: Direct-Gradient Optimization for Non-Verbal Control in Continuous Autoregressive Flow Matching Text-to-Speech
While modern text-to-speech (TTS) systems generate highly natural speech and support inline non-verbal vocalization (...
HuggingFace
Specification Before Generation: A Pre-Registered, Five-Model Paired Evaluation of a Specification Frame for LLM-Generated Code in Money, Time, Idempotency, and Access Tasks
Code generated by large language models passes security checks at a rate that has barely moved in four years. In regu...
HuggingFace
Learning from Teacher Continuations at Student States
We present OLIVE (OnLine InterVEntion). At each iteration, the evolving student policy generates a new prefix, the te...