GEMS enables multi-semantic superposition in LLMs via norm-preserving superposition, attention injection, and real-time orthogonalization, maintaining high performance on GSM8K and Wikitext-2.
Findings of the Association for Computational Linguistics: EMNLP 2025 , year =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
GEMS: Geometric Constraints Enable Multi-Semantic Superposition in LLMs
GEMS enables multi-semantic superposition in LLMs via norm-preserving superposition, attention injection, and real-time orthogonalization, maintaining high performance on GSM8K and Wikitext-2.