Blinded by Generated Contexts: How Language Models Merge Generated and Retrieved Contexts When Knowledge Conflicts?

Documenti analoghi

• Does RAG Know When Retrieval Is Wrong? Diagnosing Context Compliance under Knowledge Conflict
  di: Chen, Yihang, et al.
  Pubblicazione: (2026)
• Parameters vs. Context: Fine-Grained Control of Knowledge Reliance in Language Models
  di: Bi, Baolong, et al.
  Pubblicazione: (2025)
• ConflictRAG: Detecting and Resolving Knowledge Conflicts in Retrieval Augmented Generation
  di: Wang, Chenyu, et al.
  Pubblicazione: (2026)
• On the Consistency of Multilingual Context Utilization in Retrieval-Augmented Generation
  di: Qi, Jirui, et al.
  Pubblicazione: (2025)
• Rethinking All Evidence: Enhancing Trustworthy Retrieval-Augmented Generation via Conflict-Driven Summarization
  di: Chen, Juan, et al.
  Pubblicazione: (2025)