lesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysislesson 3 of 5 · slice-based error analysis
Checkpoint
One last thing before we move on. pass this to mark the lesson done, or skip and keep moving. hop to the next when you're ready.
Finish the analysis loop: slice, rank, then pull the reading list. Group the (segment, input_id, correct) rows, print the per-slice table worst-first (f"{s:<8} {acc:.0%} (n={n})"), then print f"worst slice: {name}" and the failing input ids from that slice as "failures to read: [...]" — because the table produces guesses and the examples produce fixes.