Same-base model merging creates a cross-domain code clone detector that reaches about 93% of multi-task performance without training data at merge time and generalizes better to AI-generated clones.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A Unified Model for Cross-Domain Clone Detection via Model Merging
Same-base model merging creates a cross-domain code clone detector that reaches about 93% of multi-task performance without training data at merge time and generalizes better to AI-generated clones.