List your company

Question Answering on LME-S

73.7Overall AccuracyC6-draw (dual-trace)
17.85232.35146.8561.349Apr 14, 2026

Evaluation results

MethodLinks
C6-draw (dual-trace)Condition=C6-draw (dual-trace)2026.0473.775508065100
C7-control (fact-only)Condition=C7-control (fact-only)2026.0453.57520552595
C5-FS (+index)Condition=C5-FS (+index)2026.0448.56020452595
C3 (baseline)Condition=C3 (baseline)2026.0447.5555553095
C4 (evidence-scored)Condition=C4 (evidence-scored)2026.0447.56010453095
C1 (broken dual-trace)Condition=C1 (broken dual-trace)2026.0440.45003520100
C2 (fact-only)Condition=C2 (fact-only)2026.0438.44015302085
Vanilla LettaCondition=Vanilla Letta2026.04200000100