Back to Report Blocks

CorrectionRate

Operations
Report block

Shows how often feedback changes AI answers, which is the inverse operational view of acceptance.

What It Answers

  • How many reviewed results were corrected?
  • Which recent items required changes?
  • How large is the feedback correction workload?

Use When

  • Operational workload reviews.
  • Comparing correction pressure between scores.
  • Finding recent corrected examples.

Avoid When

  • You need a positive stakeholder metric; use AcceptanceRate.
  • You need chance-corrected quality; use FeedbackAlignment.

Run From The CLI

Direct report commands are the simplest path for one-off usage. Saved report configurations use `plexus report config create` and `plexus report run`.

plexus feedback report correction-rate \
  --scorecard "Customer Service QA" \
  --score "Medication Review: Dosage" \
  --days 30
plexus report config create --name "CorrectionRate Example" --file correction-rate.md
plexus report run --config "CorrectionRate Example"

Minimal Configuration

class: CorrectionRate
scorecard: "Customer Service QA"
score: "Medication Review: Dosage"
days: 30
max_items: 25

Live Rendered Example

Rendered with the same dashboard component used by generated reports.

CorrectionRate Example

Feedback correction rate for selected score results.

March 1, 2026 to March 31, 2026

Corpus Correction Rate
14.17%
Items
80
Corrected Score Results
17
Total Score Results
120
Scorecard: Customer Service QA • Score: Medication Review: Dosage • Rendering 1 of 1 rows
ItemScore ResultsCorrectedUncorrectedCorrection Rate
item-00121150.00%
All rows loaded (1)

How To Interpret It

  • High correction rate means reviewers often disagree with AI outputs.
  • Pair this with contradiction analysis before assuming every correction is valid training signal.