Date
AI Overview
Why it matters
Unreliable or poorly calibrated uncertainty estimates in AI-based automatic grading can lead to unstable instructional decisions and harm student learning and equality.
Study at a glance
- Question: Which uncertainty metrics best identify unreliable grading decisions made by large language models?
- Setting and data: 14 questions across essay writing, chemistry, science, and math pedagogy; includes student answers from grades 7–10 and adult learners.
- Method: Model or benchmark evaluation testing repeated-generation uncertainty metrics across 14 LLMs and three prompting strategies.
- AI tools: Calude Haiku 4.5, Calude Sonnet 4.5, Gemini 2.5 Pro, Gemini 3 Flash, GPT-5, Llama 3.1 8B Instruct
Key findings
- Categorical-entropy consistently attained the top rank across evaluation metrics, questions, models, and prompting strategies for identifying grading errors.
- Categorical-based uncertainty measures consistently outperformed relation-based methods across tasks, although relation-based graph metrics provided greater stability across sample sizes.
- Categorical metrics worked best for lower-performing models, whereas relation-based methods became more effective on higher-performing models with more coherent rationales.
Summary drafted with AI from the full paper and may contain errors. See the full paper for details.
The rapid rise of large language models (LLMs) is reshaping the landscape of automatic assessment in education. While these systems demonstrate substantial advantages in adaptability to diverse question types and flexibility in output formats, they also introduce new challenges related to output uncertainty, stemming from the inherently probabilistic nature of LLMs. Output uncertainty is an inescapable challenge in automatic assessment, as assessment results often play a critical role in informing subsequent pedagogical actions, such as providing feedback to students or guiding instructional decisions. Unreliable or poorly calibrated uncertainty estimates can lead to unstable downstream interventions, potentially disrupting students' learning processes and resulting in unintended negative consequences. To systematically understand this challenge and inform future research, we benchmark a broad range of uncertainty quantification methods in the context of LLM-based automatic assessment. Although the effectiveness of these methods has been demonstrated in many tasks across other domains, their applicability and reliability in educational settings, particularly for automatic grading, remain underexplored. Through comprehensive analyses of uncertainty behaviors across multiple assessment datasets, LLM families, and generation control settings, we characterize the uncertainty patterns exhibited by LLMs in grading scenarios. Based on these findings, we evaluate the strengths and limitations of different uncertainty metrics and analyze the influence of key factors, including model families, assessment tasks, and decoding strategies, on uncertainty estimates. Our study provides actionable insights into the characteristics of uncertainty in LLM-based automatic assessment and lays the groundwork for developing more reliable and effective uncertainty-aware grading systems in the future.
Research Categories
What is the application?
Who is the user?
Who age?
Why use AI?
Study design
AI tools
Publication Source
Publisher
arXiv
