This market is centered on whether an open source Large Language Model (LLM) will achieve a higher Elo rating than OpenAI’s GPT-4 on the Chatbot Arena platform by the end of 2024.
Chatbot Arena utilizes a crowdsourced, randomized battle platform where user votes contribute to computing Elo ratings. This specific market will be resolved based on the Elo ratings of the models as reported by Chatbot Arena.
The link to the website: https://chat.lmsys.org/
The resolution will consider the highest Elo rating recorded for an open source LLM compared to GPT-4’s Elo rating as of December 31, 2024.
It’s important to note that only the Elo ratings will be used for determining the outcome, not considering other benchmarks like MT-Bench or MMLU scores.
The latest update of the leaderboard by December 20, 2024, will be used for the final assessment.
Update 2024-12-12 (PST) (AI summary of creator comment): - The market will consider any GPT-4 model variant present on the Chatbot Arena leaderboard, not just a specific version
An open source model needs to surpass any of the GPT-4 models listed on the leaderboard to resolve as YES
Update 2024-12-12 (PST) (AI summary of creator comment): - For a YES resolution, an open source model must surpass any of the GPT-4 model variants listed on the leaderboard
Multiple GPT-4 model variants may be present on the leaderboard, and all will be considered for comparison
1,000🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ328 | |
| 2 | Ṁ100 | |
| 3 | Ṁ68 | |
| 4 | Ṁ54 | |
| 5 | Ṁ52 |
People are also trading
@Bayesian there are multiple GPT4 models (OpenAI is constantly releasing new versions) on the LLM arena.
market is predicting whether any open source model is going to be ranked higher than any of the GPT4 models.
note that market was created long time before models like LLama and Qwen were published
@FedorShabashev yeah, so I still don't get it. If, on december 20th or 31st or whatever, the current best OpenAI model in the gpt4 family is better than all the opensource models, this resolves NO?
@Bayesian An open source model needs to surpass any of the GPT-4 models listed on the leaderboard to resolve as YES
@FedorShabashev oh ok. why isn't this resolved YES already then? It says "by december" and this already happened
@Bayesian because: " The latest update of the leaderboard by December 20, 2024, will be used for the final assessment."
@FedorShabashev so the market title could be changed to "Will an Open Source LLM Surpass any GPT-4 model in Elo Rating on Chatbot Arena on december 20, 2024"?