Paper
LLMs & Texto
Comparing Linear Probes with Mahalanobis Cosine Similarity
The Mahalanobis cosine similarity provides a theoretically grounded method for comparing linear probes that correlates strongly with out-of-distribution performance metrics.
Hugging Face · Daily Papers
·Zhuofan Josh Ying, Peter Hase
·
·▲ 1 upvotes
Este artigo está em destaque na seleção diária de papers do Hugging Face, curada pela comunidade de pesquisa em IA.
Autores: Zhuofan Josh Ying, Peter Hase, Nikolaus Kriegeskorte
- 1 upvotes da comunidade
- Temas: Mahalanobis cosine similarity, linear probes, out-of-distribution AUROC, signal-to-noise ratio, Gaussian projections, sigmoid-shaped functions
Resumo
Resumo original (em inglês), extraído do paper:
The Mahalanobis cosine similarity provides a theoretically grounded method for comparing linear probes that correlates strongly with out-of-distribution performance metrics.
