AI Paper Atlas

Research map / Large language models / AI and society

GUIDE: A Benchmark for Understanding and Assisting Users in Open-Ended GUI Tasks

Saelyne Yang, Jaesang Yu, Yi-Hao Peng, Kevin Qinghong Lin, Jae Won Cho, Yale Song, Juho Kim

CVPR 2026 · Large language models › AI and society · approach: Human-centered trust and agency · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
provide behavior state as context for help need detection (ablation)GPT-4o · GUIDE Help Need DetectionF1: 47.73 -> 90.19Improved
provide behavior state as context for help need detection (ablation)Claude-4.5-Sonnet · GUIDE Help Need Detectionaccuracy: 39.49 -> 58.56Improved
add behavior state and intent to help content prediction (ablation)InternVideo2.5-8B · GUIDE Help Content Predictionaccuracy: 23.67 -> 73.86Improved
add behavior state and intent to help content prediction (ablation)Claude-4.5-Sonnet · GUIDE Help Content Predictionaccuracy: 55.00 -> 82.79Improved
add intent on top of behavior state for help need detection (ablation)Gemini-2.5-Pro · GUIDE Help Need Detectionaccuracy: 84.73 -> 82.38Hurt
add intent on top of behavior state for help need detection (ablation)GPT-4o · GUIDE Help Need Detectionaccuracy: 87.79 -> 87.91No clear effect
add previous behavior state for behavior detection (ablation)MLLMs · GUIDE Behavior State Detectionaccuracy: 21.57 -> 27.02 (InternVideo2.5-8B)Improved
add behavior state context for intent prediction (ablation)MLLMs · GUIDE Intent Predictionaccuracy: 71.39 -> 72.62 (Claude-4.5-Sonnet)Improved
give models progressively more of the video segment (25% to 100%) (ablation)MLLMs · GUIDE online settingaccuracyImproved
use stricter MBAcc instead of accuracyMLLMs · GUIDE Intent Predictionmetric: 71.39 -> 65.44 (Claude-4.5-Sonnet)Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy