Interleaving cross-block quantization so seam pairs are refined twice reduces perplexity and prevents divergence in low-bit LLM compression, with a conditional bound showing seams multiply an error contraction factor.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
From Sweep to Seam: Interleaved Cross-Block Post-Training Quantization
Interleaving cross-block quantization so seam pairs are refined twice reduces perplexity and prevents divergence in low-bit LLM compression, with a conditional bound showing seams multiply an error contraction factor.