Skip to content

Get independent Run 003 eval from an external user #4

Description

@ardamoustafa1

Goal

Fill the proof gap: a third-party run using evals/RUN_TEMPLATE.md, raw outputs, and blind judge scores.

Acceptance criteria

  • External runner, not the maintainer
  • Same model/settings in control and treatment arms
  • Raw generated artifacts attached or linked
  • Validator logs attached
  • Three judge passes or human judge notes included
  • Results added to evals/RESULTS.md, positive or negative

Starter path

Open an issue using the Independent eval template or PR a completed evals/runs/run-003.md.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions