Lizard linearizes Transformer LLMs via subquadratic attention and adaptive learnable modules, recovering near-original performance while outperforming prior linearization methods on MMLU and associative recall.
First, Lizard still relies on a strong pretrained backbone to achieve high quality
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Lizard: An Efficient Linearization Framework for Large Language Models
Lizard linearizes Transformer LLMs via subquadratic attention and adaptive learnable modules, recovering near-original performance while outperforming prior linearization methods on MMLU and associative recall.