heterogeneous-fusion

Tag

Cards List
#heterogeneous-fusion

Training-Free Knowledge Transfer Across Model Scales through Activation-Guided Pruning

arXiv cs.LG · 2026-08-17 Cached

This paper proposes Activation-Prune-Merge (APM), a training-free framework for cross-scale fusion that improves smaller language models using larger donors without semantic alignment, achieving performance gains on multiple benchmarks.

0 favorites 0 likes
← Back to home

Submit Feedback