Cosine similarity poorly predicts performance degradation from layer removal in LLMs, making direct accuracy-drop ablation a more reliable relevance metric.
Journal of the american statistical association , volume=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Sensitivity analysis of tactical wireless network design via Tabu Search reveals scale-dependent transitions where some parameters reshape topology while others mainly scale performance magnitude.
citing papers explorer
-
Rethinking Layer Relevance in Large Language Models Beyond Cosine Similarity
Cosine similarity poorly predicts performance degradation from layer removal in LLMs, making direct accuracy-drop ablation a more reliable relevance metric.
-
Sensitivity Analysis of Tactical Wireless Network Design Under Realistic Operational Constraints
Sensitivity analysis of tactical wireless network design via Tabu Search reveals scale-dependent transitions where some parameters reshape topology while others mainly scale performance magnitude.