Model Releases
SuPER-Mamba: Surprisal Prioritization for Extended-Context Routing in Mamba
arXiv:2610.16636v1 Announce Type: cross Abstract: State space models provide efficient linear-complexity processing for long sequences, but their performance frequently degrades as context length grow
arXiv:2610.16636v1 Announce Type: cross 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.
Related
- Hidden State Poisoning Attacks against Mamba-based Language Models
- Mamba-based Selective State Space Modeling Improves the Accuracy-Complexity Tradeoff of SmolVLA Vision-Language-Action Experts
- DM3D: Dynamic Mamba via Offset-Guided Feature Resampling for Point Cloud Understanding
Source: arXiv cs.AI | 2026-10-15