set-1: 125 questions on rung-00100
  25/125
  50/125
  75/125
  100/125
  125/125
  wrote predictions-set-1.jsonl and handoff-set-1.jsonl

set-2: 124 questions on rung-00100
  25/124
  50/124
  75/124
  100/124
  wrote predictions-set-2.jsonl and handoff-set-2.jsonl

🔴 No score is computed here and none may be. Scoring is prompt 6's, from Codex.
