OPRM, a training-free chunk-and-select inference method, improves recurrent LLMs by 14-51% on LongBench and sets a 7B-class LongBench v2 record, while raising doubts about whether recurrent models exploit long-range dependencies.
Mambaextend: A training-free approach to improve long context extension of mamba
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Overflow Prevention Enhances Long-Context Recurrent LLMs
OPRM, a training-free chunk-and-select inference method, improves recurrent LLMs by 14-51% on LongBench and sets a 7B-class LongBench v2 record, while raising doubts about whether recurrent models exploit long-range dependencies.