MergeLock applies random invertible matrix transformations to Transformer attention weights, preserving the model's own output while forcing any merged model's accuracy down to near random.
Instructional fingerprinting of large language models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Model Unmerging: Making Your Models Unmergeable for Secure Model Sharing
MergeLock applies random invertible matrix transformations to Transformer attention weights, preserving the model's own output while forcing any merged model's accuracy down to near random.