Back to Report Blocks
FeedbackContradictions
Curation
Report block
Finds feedback edits that appear to contradict the current rubric, grouping likely invalid feedback into reviewable topics.
What It Answers
- Which feedback items may be invalid under the current rubric?
- What policy themes explain the contradictions?
- Which examples should be reviewed before feedback curation?
Use When
- After rubric changes.
- Before building feedback datasets.
- Before optimizing against feedback that may contain stale policy labels.
Avoid When
- You want aggregate alignment only; use FeedbackAlignment.
- You need a final invalidation action; this report proposes candidates but does not replace review.
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 contradictions \
--scorecard "Customer Service QA" \
--score "Medication Review: Dosage" \
--days 90 \
--include-rubric-memory \
--format jsonplexus report config create --name "FeedbackContradictions Example" --file feedback-contradictions.md
plexus report run --config "FeedbackContradictions Example"Minimal Configuration
class: FeedbackContradictions
scorecard: "Customer Service QA"
score: "Medication Review: Dosage"
days: 90
include_rubric_memory: true
max_feedback_items: 400Live Rendered Example
Rendered with the same dashboard component used by generated reports.
FeedbackContradictions Example
Medication Review: Dosage
9 contradictions found across 64 feedback items
Identifies feedback edits that appear to contradict the current score guidelines — so that either the feedback can be marked invalid, or the guidelines can be updated to reflect the correct policy.
14% contradiction rate
Customer acknowledgment required by stale feedback
9
How To Interpret It
- Topics are review queues, not automatic invalidation commands.
- High-confidence topics with reviewer comments are the strongest curation candidates.
- Rubric-memory citations help explain whether a correction reflects current policy or stale historical guidance.