leaderboard / docs /colosseum_top.md
Jae-Won Chung
Move doc position
08c080c

A newer version of the Gradio SDK is available: 4.36.1

Upgrade

Enter the ML.ENERGY Colosseum, where language models duel with intellect, and your judgment tips the scales of victory.

Rules of the Colosseum

  • As the spectator, you'll decide the fates of two anonymous language models -- our gladiators.
  • Your role is twofold: First, you vote for the model that delivered the best response to your prompt.
  • Next, mighty Zeus will reveal which language model consumed more energy. Evaluate if its performance justified the energy consumption.
  • Only after you cast votes will the models' identities be unveiled.

Terms of use

By using our service, you agree to these Terms of Use and accept that the model inference energy usage provided by the Service can have measurement errors. We are not liable for any damages or loss incurred by you or any third party arising from the use of the Service. It may generate offensive content and offers limited safety measures, thus should not be used for any illegal, harmful, violent, racist, or sexual purposes. The service is research purposes only. The service collects user dialogue data and voting results. We reserve the right to distribute the dataset in the future.