LLMs perform well on basic syntactic and semantic bugs in small code but struggle with complex security vulnerabilities and large production codebases.
LLM & HPC: Benchmarking DeepSeek’s Performance in High–Performance Computing Tasks
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Can LLMs Find Bugs in Code? An Evaluation from Beginner Errors to Security Vulnerabilities in Python and C++
LLMs perform well on basic syntactic and semantic bugs in small code but struggle with complex security vulnerabilities and large production codebases.