You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Great research!
Thanks for making the data and code available.
I am considering conducting research using your data and have the following question:
Is there a mapping between the topic clustering (cluster_id) to the prompt / prompt id/ conversation or anything like that?
(i saw the cluster id in the opening_prompt_cluster_df_original.csv file, but i couldn't really map it back to the conversation.)
is the moderation run on the user_prompt, and not on the model_response?
Thank you,
The text was updated successfully, but these errors were encountered:
Thanks for messaging :) You should be able to map the text id back to the cluster id with this text-level file: results/clusters/opening_prompt_text_df_original.csv
Moderation is run on every major type of text instance - model response, user_prompt, open_feedback (from conversations), as well as system_string, self_description (from survey).
Hi,
Great research!
Thanks for making the data and code available.
I am considering conducting research using your data and have the following question:
Is there a mapping between the topic clustering (cluster_id) to the prompt / prompt id/ conversation or anything like that?
(i saw the cluster id in the opening_prompt_cluster_df_original.csv file, but i couldn't really map it back to the conversation.)
is the moderation run on the user_prompt, and not on the model_response?
Thank you,
The text was updated successfully, but these errors were encountered: