MuSiQue

The 'MuSiQue' dataset/benchmark is used to evaluate the factual correctness of answers generated by large language model systems in the context of knowledge graph integration and retrieval.

Papers using MuSiQue (8)

Logit-Contribution Scoring Identifies Non-Literal Retrieval Heads2026Retrieval as Reasoning: Self-Evolving Agent-Native Retrieval via LLM-Wiki2026MeMo: Memory as a Model2026PersonalAI 2.0: Enhancing knowledge graph traversal/retrieval with planning mechanism for Personalized LLM Agents2026OCC-RAG: Optimal Cognitive Core for Faithful Question Answering2026Wikontic: Constructing Wikidata-Aligned, Ontology-Aware Knowledge Graphs with Large Language Models2025MacRAG: Compress, Slice, and Scale-up for Multi-Scale Adaptive Context RAG2025Chain-of-Thought Matters: Improving Long-Context Language Models with Reasoning Path Supervision2025