No single detector excels across all conditions; model rankings and apparent effectiveness depend heavily on the chosen test data and metrics.
Ta- ble 1 summarizes the selected model variants, de- tailingtheiroriginalevaluationdata, metrics, andre- ported performance
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Spotlights and Blindspots: Evaluating Machine-Generated Text Detection
No single detector excels across all conditions; model rankings and apparent effectiveness depend heavily on the chosen test data and metrics.