most citedSpiroLLM: Finetuning Pretrained LLMs to Understand Spirogram Time Series with Clinical Validation in COPD Reporting

1 citations · 1 across the 4 of their papers we have counts for

collaborators
Showing cs.CLShow all

2 papers · 1 filter

cs.CL2026

When Does Better Acceptance Pay? A Controlled Study of Speculative Decoding

Yongchao Long, Chunrui Zheng, Yuxuan Zhou +10

The throughput gains of speculative decoding depend on the trade-off between acceptance length and execution latency. We ask when additional acceptance is worth the cost of obtaini…

cs.CL2025

Copy-Paste to Mitigate Large Language Model Hallucinations

Yongchao Long, Xian Wu, Yingying Zhang +3

While Retrieval-Augmented Generation (RAG) enables large language models (LLMs) to generate contextually grounded responses, contextual faithfulness remains challenging as LLMs may…