Attention heads exhibit negative higher-order synergy (negative triple dividends), allowing pruning of redundant heads that cuts FLOPs by ~18% with only small perplexity increase.
Su, et al., Shrp: Specialized head routing and pruning for efficient encoder compression, arXiv preprint arXiv:2512.20635 (2025)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
baseline 1
citation-polarity summary
fields
cs.AI 1years
2026 1verdicts
UNVERDICTED 1roles
baseline 1polarities
baseline 1representative citing papers
citing papers explorer
-
A Game Theoretic Free Energy Analysis of Higher Order Synergy in Attention Heads of Large Language Models
Attention heads exhibit negative higher-order synergy (negative triple dividends), allowing pruning of redundant heads that cuts FLOPs by ~18% with only small perplexity increase.