iScore: Visual Analytics for Interpreting How Language Models Automatically Score Summaries
Open Access
- 18 March 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 787-802
Abstract
The recent explosion in popularity of large language models (LLMs) has inspired learning engineers to incorporate them into adaptive educational tools that automatically score summary writing. Understanding and evaluating LLMs is vital before deploying them in critical learning environments, yet their unprecedented size and expanding number of parameters inhibits transparency and impedes trust when they underperform. Through a collaborative user-centered design process with several learning engineers building and deploying summary scoring LLMs, we characterized fundamental design challenges and goals around interpreting their models, including aggregating large text inputs, tracking score provenance, and scaling LLM interpretability methods. To address their concerns, we developed iScore, an interactive visual analytics tool for learning engineers to upload, score, and compare multiple summaries simultaneously. Tightly integrated views allow users to iteratively revise the language in summaries, track changes in the resulting LLM scores, and visualize model weights at multiple levels of abstraction. To validate our approach, we deployed iScore with three learning engineers over the course of a month. We present a case study where interacting with iScore led a learning engineer to improve their LLM’s score accuracy by three percentage points. Finally, we conducted qualitative interviews with the learning engineers that revealed how iScore enabled them to understand, evaluate, and build trust in their LLMs during deployment.Keywords
All Related Versions
Funding Information
- National Science Foundation (2247790, 2112532)
This publication has 49 references indexed in Scilit:
- Using Large Language Models to Provide Formative Feedback in Intelligent Textbooks2023
- Visual analytics of potential dropout behavior patterns in online learning based on counterfactual explanationJournal of Visualization, 2022
- The return of intelligent textbooksAI Magazine, 2022
- Multitask Summary Scoring with Longformers2022
- Discourse synthesis: Textual transformations in writing from sourcesReading and Writing, 2022
- Visual analytics of video‐clickstream data and prediction of learners' performance using deep learning models in MOOCs' coursesComputer Applications in Engineering Education, 2020
- Automated Summarization Evaluation (ASE) Using Natural Language Processing Tools2019
- Beyond reading comprehension: exploring the additional contribution of Core Academic Language Skills to early adolescents’ written summariesReading and Writing, 2018
- Visual learning analytics of educational data: A systematic literature review and research agendaComputers & Education, 2018
- MOOC visual analytics: Empowering students, teachers, researchers, and platform developers of massively open online coursesJournal of the Association for Information Science and Technology, 2017