Model openai10

Training model: gpt-4o-2024-11-20
Inference model: gpt-4o-mini

Investigations

Performance

DatasetValidation accuracyTest accuracy
espionage0.90.9
potions0.90.6
timetravel_insurance0.750.75
titanic0.64705882352941180.5490196078431373
wisconsin0.84126984126984130.7121212121212122
wisconsin0.82539682539682540.8181818181818182