2 papers
cs.SE2026
Can LLMs Simulate Novice Programmers' Misconceptions?
Batuhan Yeltekin, Daniel Bauer
We evaluate 13 LLMs on generating, solving, simulating, and diagnosing novice programming misconceptions on code-tracing problems. While frontier models reliably perform these task…
cs.HC2026
ForkSCOPE: Charting the Agentic Garden of Forking Paths
Arjun Balaji, Batuhan Duru Yeltekin, Tian Zheng
Even with a fixed dataset and research question, data analysis involves many defensible decisions. Understanding how these choices influence the results is scientifically important…