Edit Discussion_segmenter Template

Discussion_segmenter v1

Viewer

Test Case:

Template

Preview

History

System prompt (optional) User prompt (template)

You are segmenting an NSPE Board of Ethical Review case discussion. Each paragraph is labeled [P#]; some are marked '(cites precedent)'. Classify EACH paragraph as:
  RECAP    = it narrates the facts/actors/scenario of a CITED prior case (a precedent)
  PRESENT  = it analyzes the case under review (this may legitimately cite a precedent to APPLY its holding -- that is still PRESENT)

A precedent's actors (its own 'Engineer A', 'Engineer B') belong to RECAP paragraphs. Return ONLY a JSON object mapping each paragraph number to "recap" or "present", e.g. {"0": "recap", "1": "present"}.

{{ numbered }}

Loading preview...

Version History

No version history yet. Save changes to create versions.

Saved

606 chars

Reference

Template Variables

The discussion paragraphs, one per line as [P#] (cites precedent): text

Output Schema

labels (dict[str,str])

<paragraph index>* "recap" | "present"

JSON Wrapper

Appended to the prompt by the extractor:

IMPORTANT: Wrap your response as a JSON object with exactly two keys:
{"new_discussion_segmenter_classes": [...], "discussion_segmenter_individuals": [...]}
If there are no individuals to report, use an empty array for "discussion_segmenter_individuals".