Can you say more about how this applies to long-running conversations? I've been thinking about them as well, but can't write wrap my head around how this would be better than (or even different to) standard compaction.
replystandard compactions doesnt really distinguish between long term vs short term ephemeral facts ?
replyForgive me if I'm being naive, but can't you just tweak the compaction prompt to differentiate? Presumably that's what you would do in the separate prompt anyway, right?
replyThis sounds like we are trying to add an LSTM into a transformer
replySepp would like a word
reply