Probabilistic LLM summaries show stronger semantic alignment and contextual coverage while deterministic approaches yield more concise and fully reproducible outputs in a controlled study of design pattern Java code.
2025 IEEE/ACM International Workshop on Large Language Models for Code (LLM4Code) , year =
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
-
Deterministic vs. Probabilistic Summarisation: An Empirical Trade-off Study in Design Pattern Centric Java Code
Probabilistic LLM summaries show stronger semantic alignment and contextual coverage while deterministic approaches yield more concise and fully reproducible outputs in a controlled study of design pattern Java code.