让AI评分模型适应新评分标准，提升跨标准评估能力。
When Rubrics Change: Cross-Rubric Generalization for Critical Thinking Essay Scoring
- 用中间特征表示法分离评分标准与作文内容，提升泛化性。
- 在最难场景下，宏平均F1提升5.0%，优于无特征基线。
- 适合需要灵活应对新评分体系的教育AI开发人员。
自动化作文评分（AES）研究多关注跨题型泛化，即对未见题目作文进行评分，但评分标准通常保持不变。现实中，教师常修改或新增评分标准以评估不同写作维度。本文研究跨评分标准泛化：在一种评分标准下训练，却对从未见过的新标准进行评分。采用基于大语言模型（LLM）的微调框架，包含两类组件：与评分标准无关的中间表示（称为traits），以及在训练中使用已知标准下的目标作文监督信号。在引入多个评分标准标签的批判性思维作文数据集上，实验表明，traits使最困难设置（目标标准和目标作文均未见过）下的宏平均F1提升5.0%。增加目标作文监督可进一步提升性能，最优开源Llama模型微调后比GPT-5-mini提示高出2.1%宏平均F1，仅落后GPT-5 1.9%。结果证明，基于trait的中间结构与受控监督能有效提升对未知评分标准的泛化能力。
原文摘要 · Abstract (English)
Automated essay scoring (AES) research has largely focused on cross-prompt generalization, where essays from unseen prompts are scored while the scoring criteria are typically held constant. In practice, however, educators may revise or even introduce new rubrics in their scoring task, to evaluate different aspects of essays. We study cross-rubric generalization: training on essays labeled under one set of rubrics and evaluating on previously unseen rubrics, which target different aspects of the essay. We use a Large Language Model (LLM) fine-tuning framework with two components: rubric-agnostic intermediate representations, called traits, and target-essay supervision under seen rubrics during training. On an AES dataset augmented with multiple rubric-defined labels of student critical thinking skills, we find that traits improve macro F1 by 5.0% over a baseline without traits in the hardest setting, where both target rubrics and target essays are unseen during training. We further find that increasing target-essay supervision improves performance, with our best fine-tuned open-source Llama-based model outperforming GPT-5-mini prompting by 2.1% macro F1 and trailing GPT-5 by 1.9%. These results show that trait-based intermediate structure and controlled supervision improve generalization to unseen rubrics.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。