ResearchBenchmarksAction Extraction on Human-labeled gold set 80 missionsAction Extraction on Human-labeled gold set 80 missionsFollow87.68PrecisionMAQV pipeline83.29685.48887.6889.872Mar 19, 2026Evaluation resultsMethodMethodLinksPrecisionRecallF1 (micro)Exact Sequence MatchNormalized Edit DistanceMAQV pipeline2026.0387.6889.788.2742.513.81