Md Saleh Ibtasham

Let’s read the log: root cause analysis of railway test execution logs with large language models

Results showed that long-context LLMs tended to achieve higher accuracy than smaller models, suggesting that LLMs are currently better suited to support human-in-the-loop root cause analysis than to fully automate it, and motivating further work to improve prediction accuracy for log-based RCA.