Lachesis predicts whether AutoFL's self-consistency vote will rank the true buggy method first, using LSTM and GCN models over reasoning path matrices and graphs.
Large language models for software engineering: Survey and open problems,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.SE 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Lachesis: Predicting LLM Inference Accuracy using Structural Properties of Reasoning Paths
Lachesis predicts whether AutoFL's self-consistency vote will rank the true buggy method first, using LSTM and GCN models over reasoning path matrices and graphs.