How Temperature Shapes Ideological Discourse in Retrieval-Augmented Generation?

  • 类型:arxiv
  • 标识:2607.11783
  • 链接:http://arxiv.org/abs/2607.11783v1
  • 主分类:rag
  • 形态:position
  • 被引:0
  • 被引来源:Semantic Scholar + OpenAlex
  • S2被引:0
  • OpenAlex被引:0
  • 影响力被引:0
  • TLDR:Examining the influence of the RAG framework, comprising ideological discourses, in LLM-generated answers shows that the RAG framework is prone to transferring ideological discourses into LLM responses, with sampling temperature having a measurable impact on the strength of this transfer.
  • OpenAlex ID:W7168244598
  • OpenAlex DOI:10.48550/arxiv.2607.11783
  • DOI:10.48550/arxiv.2607.11783
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://doi.org/10.48550/arxiv.2607.11783
  • OpenAlex更新:2026-07-19
  • 待LLM分类:否
  • 标题中文:温度如何塑造 RAG 中的意识形态话语?
  • TLDR中文:通过考察包含意识形态话语的 RAG 框架对 LLM 生成答案的影响,发现 RAG 框架倾向于将意识形态话语传递到 LLM 响应中,且采样温度对这种传递的强度有可测量的影响。
  • 来源文件
  • /inbox/tom/_candidates/2026-07-14-agent-rag-longcontext-candidates.json
  • [S2 enrich]
  • [OpenAlex backfill]