Showing 2026 · cs.CLShow all
3 papers · 2 filters
cs.CL2026
FlashMap: Flow Map Speculative Decoding
Jeff Shen, Zhejian Zhou, Ryan Lee +2
Speculative decoding accelerates large language model (LLM) inference by using a lightweight drafter to propose a block of tokens that the target model verifies in one parallel pas…
cs.CL2026
How Language Models Process Negation
Zhejian Zhou, Tianyi Zhou, Robin Jia +1
We study how Large Language Models (LLMs) process negation mechanistically. First, we establish that even though open-weight models often provide wrong answers to questions involvi…
cs.CL2026
Conceptual Steganography
Zhejian Zhou, Jonathan May
Language Models (LMs) emit Chains-of-Thought (CoTs) that drive much of their capability. However, the same sequence that carries useful reasoning can also covertly convey messages:…