Submit your results

Upload a zip of your submission directory: one prediction-table CSV per task (named <dataset>__<task>.csv) plus a metadata.yaml describing your method. It is validated instantly against the official test tables — a leaderboard (classification / regression / recommendation) is scored only when every task in it is present and valid, and your metadata must give a name, type, and contact email. On success a pull request is opened for maintainer review; your entry appears once it is merged. Validate locally first with python -m relbench.leaderboard <dir> — see the submission guide.

name: RelGNN                    # required — display name
type: fine-tuned               # required — fine-tuned | in-context
email: jane@example.edu        # required — recorded privately; not shown
url: https://arxiv.org/abs/... # optional — paper / project / code link
note: caveats / eval details, shown on hover   # optional