Model openailong

Training model: gpt-4o-mini
Inference model: gpt-4o-mini

Investigations

Performance

DatasetValidation accuracyTest accuracy
espionage0.750.7
potions0.70.65
southgermancredit0.65384615384615390.5398230088495575
timetravel_insurance0.750.8
titanic0.74509803921568630.6862745098039216
wisconsin0.53968253968253970.5151515151515151