MAGIC: Mixed-Granularity Agent Graphs via Incremental Construction with Dense-Reward Reinforcement LearningResearchSep 23, 2026
Train Where the Quantized Model Goes: On-Policy Distillation for Low-Bit ReasoningResearchSep 23, 2026
Towards Hierarchical GNNs for multi-grid power flow: generalization across operating scenariosResearchSep 23, 2026
S$^4$R: Selective Sampling, Subspaces, and Sparse Reconstruction for Compressed Long-Context KV CachingResearchSep 23, 2026
ROBE: Reversed-Order-Biased-Experts for Extracting Extreme Long-tail Events from Historical TextsResearchSep 23, 2026
Compositional Failure in Audio-Visual LLMs: Late-Layer Prior Dominance Under Cross-modal ConflictResearchSep 23, 2026
Hy-MultiTurn: A Six-Dimensional Benchmark for Deep Multi-Turn Dialogue UnderstandingResearchSep 23, 2026
LiLiCorr: Lightweight Likelihood Correlation of Parallel Drafts for Speculative DecodingResearchSep 23, 2026