Replies: 3 comments
-
|
OK @rzmk , we'll include the clarification in the next release - v13.1.0 - which I'm pushing to release this weekend. |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
@rzmk I added a link to this discussion on the v14.0.0 release notes. |
Beta Was this translation helpful? Give feedback.
0 replies
-
|
And I also added the same footnote to the 13.0.0 release. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
@jqnatividad from the v13.0.0 pre-release notes:
I'd suggest clarifying and also improving the initial label of "Local Data" if the data is being sent to an LLM provider such as Anthropic. The "statistical metadata" can definitely include "raw data", such as entire string column values for
minandmaxfromqsv stats. I'm not too concerned about the specifics of how much or little data should be sent as that's per each person's usage but I'd rather not have a label "Local Data" and "WITHOUT sending raw data, only sending statistical metadata to Claude" along with other phrasing as those with data privacy and training concerns would have to look further to understand better exactly what data would be sent.If I have NYC 311 data the stats output may look like this (click here):
If a dataset had PII imagine the
maxandmincolumn values or other cases we may not realize.If I have NYC 311 data the frequency output may look like this (click here):
So overall I'd suggest adding at least a disclaimer maybe and phrasing that more accurately represent the implications of using the new feature.
Beta Was this translation helpful? Give feedback.
All reactions