compass-verifier-unified-llm-evaluation

【数据与研究】Lightweight verifier model for evaluating LLM outputs across multiple domains, handling diverse answer types through meta-error pattern analysis.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准