AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Probing BERT in Hyperbolic Spaces

Boli Chen, Yao Fu, Guangwei Xu, Pengjun Xie, Chuanqi Tan, Mosha Chen, Liping Jing

ICLR 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Multilingual language representation pretraining · 19 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Use a Poincaré rather than Euclidean probeBERTbase7 dependency-distance probing · Penn TreebankUUAS: 79.8 -> 83.7Improved
Use a Poincaré rather than Euclidean probeBERTbase7 dependency-distance probing · Penn TreebankDSpr.: 0.85 -> 0.88Improved
Use a Poincaré rather than Euclidean probeBERTbase7 dependency-depth probing · Penn TreebankRoot %: 88.0 -> 91.3Improved
Use a Poincaré rather than Euclidean probeBERTlarge16 dependency-distance probing · Penn TreebankUUAS: 81.7 -> 85.9Improved
Use a Poincaré rather than Euclidean probeBERTlarge16 dependency-distance probing · Penn TreebankDSpr.: 0.87 -> 0.90Improved
Use a Poincaré rather than Euclidean probeBERTbase9 sentiment probing · Movie Reviewaccuracy: 81.7 -> 84.9Improved
Fix rather than train the class meta embeddings (ablation)Poincaré sentiment probe · Movie Review, BERTbase10accuracy: 84.2 -> 84.5Improved
Fix rather than train the class meta embeddings (ablation)Poincaré sentiment probe · Movie Review, BERTbase9accuracy: 84.9 -> 84.2Hurt
Fix rather than train the class meta embeddings (ablation)Euclidean sentiment probe · Movie Review, BERTbase9accuracy: 81.7 -> 78.4Hurt
Fix rather than train the class meta embeddings (ablation)Euclidean sentiment probe · Movie Review, BERTbase10accuracy: 83.5 -> 79.1Hurt
Use hyperbolic rather than Euclidean geometrysentiment probing of Linear embeddings · Movie Reviewaccuracy: 48.4 -> 48.4No clear effect
Use curvature -0.1 instead of -1 (ablation)2d Poincaré syntax probe · Penn Treebank, BERTbase7UUAS: 40.5 -> 38.7Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy