🧪 Test?View on arXiv
Watch Where You Linearize: Predicting and Correcting Model-Merging Interference
Not provided in the abstract
model mergingfine-tuninginterference predictionoptimization
2611.25000
Builder Relevance
2d ago80%
Abstract
The paper presents a method to predict and correct interference in model merging using directional derivatives.
Reality Card
Core Claim
The first-order directional derivative of task A's outputs can predict interference in model merging, improving accuracy and identifying problematic examples.
Method / Result
LR-Merge improves all tested base mergers, with gains generally increasing as more tasks are merged.
Limitations
The predictive capability of the derivative is limited at the shared pretrained backbone and may not generalize across different tasks.
Paper to code
Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.
No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.