[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Non-LLM example where we do not in practice use original training data



>>>>> "Stefano" == Stefano Zacchiroli <zack@debian.org> writes:

    Stefano> On Mon, May 05, 2025 at 02:13:58PM -0700, Russ Allbery wrote:
    >> However, I am very leery about extending that exception to cases
    >> where people are intentionally creating that situation by
    >> deleting the input data on purpose.

    Stefano> I agree with you on this.

FWIW, I also agree.  I let the rules lawyer side of me get ahead of
everything else.
I am collecting my thoughts and will put them together into a summary
with my thinking updated based on Russ's input and Ansgar's input after
I come up with some concise to say.
I just wanted to respond to this point because in retrospect the
position I took yesterday on intentionally deleting training data was
not well-considered.



Reply to: