List your company

Truthful Question Answering on TruthfulQA (accuracy, intervention rate)

70AccuracyAligned base model
38.846.95563.1Apr 22, 2026

Evaluation results

MethodLinks
Aligned base modelModel Pair=Q→G, Pair Family=Cross-Family2026.0470—
Aligned base modelModel Pair=L→G, Pair Family=Cross-Family2026.0470—
Aligned base modelModel Pair=G→G, Pair Family=Within-Family2026.0470—
Aligned base modelModel Pair=Q→L, Pair Family=Cross-Family2026.0467—
Aligned base modelModel Pair=G→L, Pair Family=Cross-Family2026.0467—
Aligned base modelModel Pair=L→L, Pair Family=Within-Family2026.0467—
Base ModelModel Pair=Q→L, Pair Family=Cross-Family2026.0458—
Base ModelModel Pair=G→L, Pair Family=Cross-Family2026.0458—
Base ModelModel Pair=L→L, Pair Family=Within-Family2026.0458—
NUDGINGModel Pair=L→L, Pair Family=Within-Family2026.045120
BlendInModel Pair=L→L, Pair Family=Within-Family2026.045120
BlendInModel Pair=Q→L, Pair Family=Cross-Family2026.045031.9
BlendInModel Pair=G→L, Pair Family=Cross-Family2026.045019.2
Guidance ModelModel Pair=L→G, Pair Family=Cross-Family2026.0450—
Guidance ModelModel Pair=L→L, Pair Family=Within-Family2026.0450—
BlendInModel Pair=L→G, Pair Family=Cross-Family2026.044917.2
Guidance ModelModel Pair=Q→L, Pair Family=Cross-Family2026.0448—
NUDGINGModel Pair=Q→L, Pair Family=Cross-Family2026.044831.9
Guidance ModelModel Pair=Q→G, Pair Family=Cross-Family2026.0448—
NUDGINGModel Pair=L→G, Pair Family=Cross-Family2026.044717.2
Guidance ModelModel Pair=G→L, Pair Family=Cross-Family2026.0446—
Guidance ModelModel Pair=G→G, Pair Family=Within-Family2026.0446—
NUDGINGModel Pair=G→L, Pair Family=Cross-Family2026.044519.2
NUDGINGModel Pair=G→G, Pair Family=Within-Family2026.044320.9
BlendInModel Pair=G→G, Pair Family=Within-Family2026.044320.9
BlendInModel Pair=Q→G, Pair Family=Cross-Family2026.044231.5
Base ModelModel Pair=Q→G, Pair Family=Cross-Family2026.0440—
NUDGINGModel Pair=Q→G, Pair Family=Cross-Family2026.044031.5
Base ModelModel Pair=L→G, Pair Family=Cross-Family2026.0440—
Base ModelModel Pair=G→G, Pair Family=Within-Family2026.0440—