Skip to content

Base evals: GPT-OSS-20B full re-run with reasoning-channel fallback #2

Description

@CathrineShalby

Re-run GPT-OSS across core-3 (CyberMetric/SecBench/SecEval), RedSage x5 + SECURE x2, and the wiki open-ended set with the reasoning_content fallback in the harness Model. Archive pre-fallback results and report the delta the fallback recovers (~1.25% of answers were empty-content before).

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions