Shallow Prefill, Deep Decoding: Efficient Long-Context Inference via Layer-Asymmetric KV Visibility Paper • 2605.06105 • Published 5 days ago • 1
Latent Self-Consistency for Reliable Majority-Set Selection in Short- and Long-Answer Reasoning Paper • 2508.18395 • Published Aug 25, 2025
SKIML/Bllossom_llama-3.2-Korean-Bllossom-3B_srcKorean_trgEnglish_segment_5e-05_4_32_window12 Text Generation • Updated Nov 21, 2025