PACE: Towards Surfacing Hidden Conflicts in User Requests

  • 类型:arxiv
  • 标识:2609.03293
  • 链接:https://arxiv.org/abs/2609.03293
  • 主分类:rag
  • 形态:method
  • 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:PACE is introduced, a dataset for evaluating whether models can identify latent constraints, expressed as egocentric knowledge or events, that render seemingly reasonable user requests inappropriate, and PaceMaker, a multi-agent framework in which specialized agents coordinate across query reformulation, multi-hop graph traversal, and conflict-aware filtering to retrieve contextually decisive evidence.
  • OpenAlex ID:W7208738978
  • OpenAlex DOI:10.48550/arxiv.2609.03293
  • DOI:10.48550/arxiv.2609.03293
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://doi.org/10.48550/arxiv.2609.03293
  • OpenAlex更新:2026-09-06
  • 待LLM分类:否
  • 标题中文:PACE:揭示用户请求中的隐性冲突
  • TLDR中文:提出 PACE 数据集,用于评估模型能否识别以自我中心知识或事件形式表达的潜在约束(这些约束使看似合理的用户请求变得不当),以及 PaceMaker 多 Agent 框架,其中专门 Agent 通过查询重构、多跳图遍历与冲突感知过滤进行协调,以检索上下文决定性证据。
  • 来源文件:
  • /inbox/tom/_candidates/2026-09-04-agent-rag-longcontext-candidates.json
  • [OpenAlex backfill]
  • [S2 enrich]