CSR compresses LLM key-value caches through sparse dictionary coding, reaching 1 bit per cache value while roughly matching 2-bit and 4-bit quantization baselines on LongBench.
Furthermore, the loss function value after conver- gence is also reduced
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
-
CSR:Achieving 1 Bit Key-Value Cache via Sparse Representation
CSR compresses LLM key-value caches through sparse dictionary coding, reaching 1 bit per cache value while roughly matching 2-bit and 4-bit quantization baselines on LongBench.