AI 论文 · 2026-09
浏览 2026-09 发布的AI 论文内容，第 8 页，共 741 条。
- Repo-To-Skill: Distilling GitHub Repositories Into AI4AI Skills
- Language Models Can Control Their Own Attention
- Debias-SparseGPT: Bias-Aware Pruning for Large Language Models
- Percolation Dynamics in Optimization : Variance Cascades and Discrete Scale Invariance
- The Missing Temporal Link: Temporal Context Routing for Script-Driven Audio-Video Generation
- PaperCompiler: Faithful Paper-to-Code Generation via Repository-Level Specification Compilation
- Let Confidence Change, Not the Prediction: Prediction-Preserving Repair for Post-hoc Calibration
- Exploring Collaboration between a language and a non-language agent
- Enoki: Efficient Multi-Level Hallucination Detection
- Sparse Readout Prism: Explaining Logit-Lens Scores in Features Instead of Tokens
- CRISP: Cliff-awaRe Input-adaptive Sparse Prefilling with Structural-Mass-Motivated Routing
- ExecRetrieval: Measuring the Functional-Correctness Gap in Code-Embedding Retrieval
- Agent Memory Is a Surface for Endogenous Authorization Laundering
- Kirin: Animal Motion Generation from In-the-Wild Video
- ZipTok3D: High-Fidelity 3D Tokenization with Compact Token Prefixes
- Uncovering Understanding-Generation Synergy in Native Unified Multimodal Models: From Representation, Task to System
- StudentSim: Training LLM-based Student Simulators
- From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix
- H3-World: Turning Language Understanding into World Control
- Knowledge Distillation During Mid-Training Favors Reasoning over Factual Recall
- TempCloze: Can Video-LLMs Identify the Missing Middle?
- LatentPress: Context Compression Beyond Text and Vision
- Harness-of-Harness: Multi-Day Autonomous Software Development with Continual Improvement
- Does Imitation Learning Preserve Temporal Robustness in Dexterous Manipulation? An Expert-Learner Comparison Across Task Execution Speeds
- HarnessDev: Can LLMs Create and Evolve Their Own Agent Harness?
- Learning Sparse Decision Trees via Transformer Variational Auto-Encoders
- Evaluating Multimodal LLMs as Generalist Vision-Language-Action Agents for Drone Control: Commanding, Approaching, Tracking and Searching
- SMELT: Scaling Laws for Compute-Matched MoE Looped Transformers
- EmbodiedSkills: A Unified Framework for Orchestrating, Training, and Deploying VLA Agents
- On the Design Fundamentals of Pixel Text Representation Learning
- ReFlowSET: Representation-Aligned Latent Flow Matching for SAR-to-EO Image Translation
- Replacing Training with Memory: Listwise Selection for Text-to-SQL
- StudyBench: Can Self-Evolution Squeeze Textbooks for Olympiad Capability?
- DiagEvo: Diagnosis-Guided Self-Evolution via Hierarchical Error Memory
- DramaChain Bench: An End-to-End Benchmark for Short-Drama Generation
- It Takes Two to Match: Co-Evolving Generative Retriever with Reinforcement Learning
- Control-Data Flow Separation: Stable Prompt Optimization in Multi-Agent LLMs
- A Glance Is All You Need: Single-Pass Fine-Grained Image Captioning with SimLoss
- EM^2Mem: Event-Centric Multimodal Memory for Large Language Models
- The Safeguard Worked. Is the LLM System Safer?
- Autoregressive Mosaics: Probing 2D Spatial Reasoning in Text-Only Language Models