SentTrack applies LLM summarization, UMAP+HDBSCAN clustering, and the ABCDE interaction framework to GitHub issues, reporting 49% stagnation and 13% resolution rates in one repository as evidence of a dominant resolution-gap bottleneck.
Leveraging large language models to identify conversation threads in collaborative learning,
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
-
SentTrack: Sentiment-Driven Bottleneck Detection in GitHub Issue Repositories
SentTrack applies LLM summarization, UMAP+HDBSCAN clustering, and the ABCDE interaction framework to GitHub issues, reporting 49% stagnation and 13% resolution rates in one repository as evidence of a dominant resolution-gap bottleneck.