Rubric
In LLM evaluation, a scoring standard that provides an explicit, actionable definition of acceptable output, including scoring dimensions, judgment criteria and positive and negative examples.
Timeline (last 90 days)
An eval (Eval) consists of three core components: the input prompt, the execution environment, and the evaluation criteria (Rubric).
PrismaCoder is an open-source project aimed at improving multimodal code intelligence by decomposing code quality into verifiable scoring criteria (Rubrics) to unify data filtering and reinforcement learning rewards.
Writing scoring rules into a prompt in natural language has drawbacks: wording variations across different calls may cause the model's understanding to drift, and the scoring logic cannot be directly parsed or reused by programs.
All Facts (4)
吴恩达建议使用Rubric（评分标准）作为提示词中的结构化约束技术，强制AI在特定框架内进行客观分析
90%UnverifiedAn eval (Eval) consists of three core components: the input prompt, the execution environment, and the evaluation criteria (Rubric).
50%UnverifiedPrismaCoder is an open-source project aimed at improving multimodal code intelligence by decomposing code quality into verifiable scoring criteria (Rubrics) to unify data filtering and reinforcement learning rewards.
50%UnverifiedWriting scoring rules into a prompt in natural language has drawbacks: wording variations across different calls may cause the model's understanding to drift, and the scoring logic cannot be directly parsed or reused by programs.
50%