Story
arxiv_llm_reliability · Sep 2, 2026 · paper
Source brief
Loom: Weaving Diagnostic Strands into Free-Text Consensus via Embedding-Space Reweighting
arxiv.orgSep 2, 2026
original source linked
In brief
Aggregating noisy, conflicting textual hypotheses into a reliable consensus is a fundamental challenge when deploying NLP systems in real-world industrial settings. While monolithic Large Language Model (LLM) agents o...
Feed lens
agenticeval