Showing 2026Show all
3 papers · 1 filter
cs.CL2026
RouteSparse: Input-Conditional Pattern Routing for Budgeted Long-Context Prefilling
Chao Zhang, Yifan Ji, Ziyan Zhang +2
Dynamic sparse attention can reduce the quadratic cost of long-context prefilling without changing model weights. MInference assigns each attention head one pattern offline and est…
cs.CL2026
Stepwise Reasoning Enhancement for LLMs via External Subgraph Generation
Xin Zhang, Yang Cao, Baoxing Wu +2
Large language models have shown strong performance in natural language generation and downstream reasoning tasks, but they still struggle with logical consistency, factual groundi…
cs.CL2026
SGR: A Stepwise Reasoning Framework for LLMs with External Subgraph Generation
Xin Zhang, Yang Cao, Baoxing Wu +2
Large Language Models (LLMs) have demonstrated strong capabilities across diverse NLP applications, such as translation, text generation, and question answering. Nevertheless, they…