← Search

Xu Yuanchen

2 accepted papers

2025

Reverse Modeling in Large Language Models

NAACL 2025short

Humans are accustomed to reading and writing in a forward manner, and this natural bias extends to text understanding in auto-regressive large language models (LLMs). This paper investigates whether LLMs, like humans, struggle with reverse modeling, specifically with reversed text inputs. We found t…

Cited by 0SourcePDFScholar
2024

GliDe with a CaPE: A Low-Hassle Method to Accelerate Speculative Decoding

ICML 2024poster

Speculative decoding is a relatively new decoding framework that leverages small and efficient draft models to reduce the latency of LLMs. In this study, we introduce GliDe and CaPE, two low-hassle modifications to vanilla speculative decoding to further improve the decoding speed of a frozen LLM. S…

Cited by 19SourcePDFScholar