在 HLE 多选题上实测 Jev 约 19.7%

On HLE's 513 text-only multiple-choice questions (the set used in the AA Decision Index), I measured Jev at 19.7% (that's a bit below the AA chart's 21: setups differ slightly and its outputs aren't deterministic) On this exact set, always picking the longest answer scores 20.7%, vs 16.4% for random guessing That's above every decision model tested (the effect comes mostly from Math and Physics questions)