Model anthropic

Training model: claude-3-5-haiku-20241022
Inference model: gpt-4o-mini

Investigations

Performance

DatasetValidation accuracyTest accuracy
espionage0.850.8
potions0.850.7
southgermancredit0.51923076923076930.5221238938053098
timetravel_insurance0.90.75
titanic0.60784313725490190.6274509803921569
wisconsin0.79365079365079360.8333333333333334