Papers by Carolyn Jane Anderson
Substance Beats Style: Why Beginning Students Fail to Code with LLMs (2025.naacl-long)
Copied to clipboard
| Challenge: | Existing work shows that beginners struggle to prompt LLMs to solve text-to-code tasks. |
| Approach: | They propose to use a causal intervention experiment on technical vocabulary to test whether students lack the technical vocabulary needed to write good prompts and to analyze graphs that abstract how students edit prompts. |
| Outcome: | The proposed model improves student-LLM communication by predicting student failures and predicting the information content of prompts. |
GlyphPattern: An Abstract Pattern Recognition for Vision-Language Models (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing benchmarks for abstract pattern recognition are easier because they do not involve a natural language description of the pattern. |
| Approach: | They present a dataset that pairs human-written descriptions of visual patterns with three visual presentation styles. |
| Outcome: | The proposed benchmark pairs human-written and human-verified patterns with three visual presentation styles. |