When Rubrics Change: Cross-Rubric Generalization for Critical Thinking Essay Scoring
For educators and AES systems, this work addresses the practical need to adapt to evolving scoring rubrics without retraining from scratch.
This paper introduces cross-rubric generalization for automated essay scoring, where models trained on essays labeled under one set of rubrics must score essays under unseen rubrics. Using a fine-tuned LLM with rubric-agnostic traits, they achieve a 5.0% macro F1 improvement over a baseline in the hardest setting, and their best model outperforms GPT-5-mini prompting by 2.1% macro F1.
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.