conftrace_

anisotropic representation

2 papers

Explore in graph

Co-occurring keywords · papers mentioning both

isotropy restoration (1) massive activation (1) domain-critical dimension (1) semantic detector (1) domain specialization (1) activation steering (1) transformer language model (1) embedding space (1)

Papers

Embracing Anisotropy: Turning Massive Activations into Interpretable Control Knobs for Large Language Models ACL 2026 Too Much in Common: Shifting of Embeddings in Transformer Language Models and its Implications NAACL 2021