I’m considering entering DataCrunch Equity Market Neutral #2 as an individual.
May I use ChatGPT/Codex to write and improve my submission code, while running that code in my own participant-controlled environment and keeping the competition dataset and credentials out of the AI’s context?
I would share only permitted aggregate evaluation results and data-free debugging information. Does this arrangement comply with sections 5 and 6, particularly the restrictions on private code sharing and the original-work requirement? Are there any additional conditions?
You are allowed to use an LLM to help you write your model. However, be mindful when sharing data with it. You may share preview snippets or aggregation results (sum, std, avg, etc.) as it needs this information to help you, but not the full dataset.
I also suggest you always click “Open in…” on documentation pages, as we already included a lot of details that could help him (your LLM):
