Lu Sun, Xiangyi Zhang, Xiangyang Zhu, Kaiyuan Ji, Zijian Chen, Qiang Hu, Zicheng Zhang, Yuan Tian, Guangtao Zhai · 2026 · arXiv (Cornell University)
DOI: 10.48550/arxiv.2610.26787 · cited by 0 works (OpenAlex) · open-access PDF
Os benchmarks de avaliação tradicionais reduzem o conhecimento científico intrinsecamente interligado nas ciências da vida a listas planas de perguntas, ignorando a estrutura topológica subjacente do conhecimento. Introduzimos, o primeiro benchmark gráfico-estruturado para as ciências da vida, com mais de 10.000 perguntas de alta qualidade de múltipla escolha em medicina, biologia e química. Nossa abordagem constrói um gráfico de avaliação ponderado usando algoritmos de correspondência bidirecional e semântica, onde os nós representam perguntas e os pesos de borda capturam suas relações semânticas. Utilizando esta topologia gráfica, projetamos duas novas métricas de avaliação. O Global Coherence Score (GCS) mede a consistência de um modelo dentro de bairros semanticamente relacionados, enquanto o
Machine translation (Meta M2M100); the English original follows.
Traditional evaluation benchmarks reduce inherently interconnected scientific knowledge in life sciences into flat lists of questions, disregarding the underlying topological structure of the knowledge. We introduce, the first graph-structured benchmark for life sciences, featuring over 10,000 high-quality multiple-choice questions across medicine, biology, and chemistry. Our approach constructs a weighted evaluation graph using bidirectional matching and semantic similarity algorithms, where nodes represent questions and edge weights capture their semantic relationships. Leveraging this graph topology, we design two novel evaluation metrics. The Global Coherence Score (GCS) measures a model's consistency within semantically related neighborhoods, while Knowledge Balance Score (KBS) analyzes how model errors are distributed across the graph to reveal conceptual blind spots. LG-Bench facilitates fine-grained comparison of LLMs by surfacing differences in conceptual coherence and patterns of knowledge organization across models. Our framework shifts the evaluation paradigm from flat accuracy metrics to structure-aware analysis, offering a new lens for diagnosing and improving LLM performance in the life sciences domain.
Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese
Topics: Computer science, Benchmark (surveying), Consistency (knowledge bases), Graph, Coherence (philosophical gambling strategy), Matching (statistics)
GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.26787, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txtSources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange