Thanks for the article and all you do as an organization.
It would be very nice to have access to the raw data, code that generated the graphs etc... in the tradition of reproducible science.
Leading by example, I always endeavor to release raw data and source for analysis and graphs. I don't always succeed. Here are some examples of notebooks backing Reddit posts--very small but not that much different:
https://www.reddit.com/r/LocalLLaMA/comments/1kd68gz/impact_of_schema_directed_prompts_on_llm/
https://www.reddit.com/r/LocalLLaMA/comments/1k65lqd/experiment_can_determinism_of_llm_output_be/
Once you get in the habit of doing this I think you will find it quite natural.
Again, thanks for the work.
