MMER merges multiple multimodal LLMs by averaging task vectors and applying per-modality binary masks, retaining about 99% of original task performance without additional training.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Multi-Modality Expansion and Retention for LLMs through Parameter Merging and Decoupling
MMER merges multiple multimodal LLMs by averaging task vectors and applying per-modality binary masks, retaining about 99% of original task performance without additional training.