Search-QA

Search-QA is a benchmark dataset for evaluating AI models on question-answering tasks, developed by researchers at Google. It tests a model's ability to retrieve and synthesize answers from web search results, emphasizing factual accuracy and compreh

Two-hop subgraph: this entity, every entity it directly relates to, and every entity those neighbors relate to. Drag a node, scroll to zoom, click to inspect — or click any neighbor and re-center the atlas there.