Search papers, labs, and topics across Lattice.
University of Passau
3
0
3
Most enrichment strategies in RAG systems backfire, with a smaller model outperforming a larger one by 19 F1 points when metadata aligns with model capabilities.
Local-first information retrieval can achieve cloud-level answer quality while keeping sensitive data on user devices, redefining the tradeoff between search scope and quality.
Trust calibration in AI-generated content can be improved by 29% using a novel user simulator that accounts for verification effort and source credibility.
