← Search

Timur Carstensen

2 accepted papers

2026

Selective Rotary Position Embedding

ICLR 2026poster

Position information is essential for language modeling. In softmax transformers, Rotary Position Embeddings (\textit{RoPE}) encode positions through \textit{fixed-angle} rotations, while in linear transformers, order is handled via input-dependent (selective) gating that decays past key-value assoc…

Cited by 0SourceScholar
2025

DeltaProduct: Improving State-Tracking in Linear RNNs via Householder Products

NeurIPS 2025poster

Linear Recurrent Neural Networks (linear RNNs) have emerged as competitive alternatives to Transformers for sequence modeling, offering efficient training and linear-time inference. However, existing architectures face a fundamental trade-off between expressivity and efficiency, dictated by the stru…

Cited by 0SourcecodeScholar