2 papers
cs.CL2026
APCD: Adaptive Path-Contrastive Decoding for Reliable Large Language Model Generation
Tianyu Zheng, Hong Wu, Jiaji Zhong
Large language models (LLMs) often suffer from hallucinations due to error accumulation in autoregressive decoding, where suboptimal early token choices misguide subsequent generat…
cs.CL2026
ReST-KV: Robust KV Cache Eviction with Layer-wise Output Reconstruction and Spatial-Temporal Smoothing
Yongqi An, Chang Lu, Kuan Zhu +5
Large language models (LLMs) face growing challenges in efficient generative inference due to the increasing memory demands of Key-Value (KV) caches, especially for long sequences.…