LLM-generated code matches human-written code in overall readability but exhibits different issue patterns, and prompt engineering has limited impact on improving it.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
The Readability Spectrum: Patterns, Issues, and Prompt Effects in LLM-Generated Code
LLM-generated code matches human-written code in overall readability but exhibits different issue patterns, and prompt engineering has limited impact on improving it.