SE-Attn and HyLoRA fine-tune hybrid SSMs on sequences up to 8x the pre-training length, approaching full-attention performance at lower cost.
Longlo RA : Efficient fine-tuning of long-context large language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Expansion Span: Combining Fading Memory and Retrieval in Hybrid State Space Models
SE-Attn and HyLoRA fine-tune hybrid SSMs on sequences up to 8x the pre-training length, approaching full-attention performance at lower cost.