Truthfulness Classification on TruthfulQA (held-out evaluation) benchmark leaderboard · SOTA2 Research

Truthfulness Classification on TruthfulQA (held-out evaluation)

82AccuracyReal-time Hook

Evaluation Results

Method	Links
Real-time Hook2026.03		82	86.25	90.55
Baseline2026.03		81.33	81.25	91.59
Post-hoc Mean2026.03		81.33	81.25	91.57
Post-hoc Pct802026.03		81.33	81.25	91.59
Post-hoc Amplify2026.03		81.33	81.25	91.55
Post-hoc Fourier2026.03		81.33	81.25	91.59
Post-hoc Zero2026.03		81.33	81.25	91.57
Real-time Hook2026.03		80	73.75	90.27
Baseline2026.03		79.33	73.75	90.86
Post-hoc Mean2026.03		79.33	73.75	90.8
Post-hoc Pct802026.03		79.33	73.75	90.86
Post-hoc Amplify2026.03		79.33	73.75	90.86
Post-hoc Fourier2026.03		79.33	73.75	90.86
Post-hoc Zero2026.03		79.33	73.75	90.86
Real-time Hook2026.03		71.33	78.75	80.91
Baseline2026.03		69.33	71.25	78.59
Post-hoc Mean2026.03		69.33	70	78.36
Post-hoc Pct802026.03		69.33	70	78.54
Post-hoc Amplify2026.03		69.33	70	78.38
Post-hoc Fourier2026.03		69.33	71.25	78.64
Post-hoc Zero2026.03		69.33	70	78.52