arXiv preprint ·
SuPER-Mamba: Surprisal Prioritization for Extended-Context Routing in Mamba
- Published
- Source
- arXiv preprint, posted before peer review
- Original
- Read it at arXiv
- Topics
Abstract
State space models provide efficient linear-complexity processing for long sequences, but their performance frequently degrades as context length grows. We propose SuPER-Mamba, a training-free framework that uses Surprisal as a model-intrinsic signal for long-range state routing in pretrained Mamba models. SuPER-Mamba computes position-wise Surprisal from next-token predictions, aggregates neighboring responses through Local Pooling, and uses the resulting scores to prioritize Global-Channel updates without query-dependent scoring or additional training. We evaluate SuPER-Mamba on three state-space models across the LongBench-E benchmark. The results show that SuPER-Mamba improves the average score from 8.06 to 15.69, corresponding to a 94.7% relative gain over the Base Models and a 13.1% improvement over the state-of-the-art prior baseline. Ablation and length-wise analyses further support the effectiveness of the proposed routing framework.
The authors’ abstract, as arXiv lists it. arXiv’s listings, abstracts included, are free for anyone to reuse.
Read it on ReadSearch
Each opens after you create an account, and brings you back to this paper.
More in Computation & Language / NLP
PAIR: Memory-Efficient and Fast Inference with Reading--Reasoning Fusion
Everyone Was Right: Lexical Representations in Language Models Are Both Abstract and Holistic
Improving the Evaluation of Biomedical Question Answering with Few Human Preference Signals
Many Steps into One: Reasoning Contraction and the Limits of Safe Memory Reuse for LLM Agents
Of Errors and Echoes: Responsibility-Aware Credit Assignment for Tool-Use Agents
Get the few papers worth your time in Computation & Language / NLP every morning.
ReadSearch reads the new papers in your field each day and brings you the few worth reading, each with its key finding in a sentence.