Get Started
Home
Topics
Search
Library
Research questionDoes Cantonese-specific adaptation improve language-model prediction of naturalistic Cantonese reading consistently across information-theoretic measures?Language models estimate information-theoretic expectations that may help explain human reading, but it is unclear whether Cantonese-specific adaptation improves this alignment. Different measures can rank models differently, making conclusions sensitive to how predictive fit is quantified.
AI
Evaluation & Benchmarks
LLM Pretraining & Post-training
Natural Language Processing
Research Paper
Latest papersRecent research connected to this question, newest first.Do Cantonese-Adapted Language Models Better Predict Cantonese Reading? A Cross-Model Eye-Tracking EvaluationThe evidence comes from naturalistic Cantonese eye-tracking data and compares two within-family adaptation contrasts: CKIP GPT-2 Tiny with JED351, and Qwen2.5-7B with CantoneseLLM-7B. It evaluates lexical surprisal, POS surprisal, entropy before the target, and entropy reduction; the reported rankings vary by measure, so the findings establish metric-dependent predictive fit rather than a general model-quality result.research paper · Sep 7, 2026
Related questions
How can models adapt to low-resource languages without damaging source-language and related-task performance?Do newer, larger vision-language models reliably improve autonomous-driving performance without task-specific adaptation?How can we compare language models’ conditional behavior and predict the effects of prompt changes?How can language-model attention remain reliable beyond its training context?