Personality

Discuss specifics of personality design, including what Keyphrases work well and what dont, use of plug-ins, responses, seeks, and more.

Posts 502 - 513 of 5,105

22 years ago #502
Unless I've used it without realizing it...no picking on me! I'm new to this! It generally makes sense, unless someone starts using long confusing words against her...

22 years ago #503
Who's picking on you?

22 years ago #504
well I change the rankings if some responses aren't getting used much but that's a fairly small percentage of my responses. And I'm not claiming she always makes sense but her answers usually have something to do with what's said.

22 years ago #505
*evil glare* Nah, I honestly never understood what the ranking thing was for. I can be slow like that.

22 years ago #506
read the book of ai... your bot should even become better

22 years ago #507
Read it before...but I shall read it now anyhow.

22 years ago #508
Lunar- try "My Bots" now, you should be able to see Gaia's Transcript.

I havent pruned Gaia in quite some time so I expect it's quite a mess. One of these days I'll tune it up a bit.

22 years ago #509
thanks doc

22 years ago #510
VALUABLE PERSONALITY ADVICE
------------------------------
1) Keyphrase Specificity: In order to get the most out of your Keyphrases, try not to be too specific. For example, the Keyphrase "do you" will match a LOT of incoming messages, while the Keyphrase "do you like to eat bars of soap" will probably never match an incoming message. You need a good balance between specificity and generality. "do you like" and "do you like to" are also good Keyphrases, but "do you like pears" and "do you like to swim" are not.
2) Bland Responses: Though bland responses like "Ok." and "Nahh.." and "Really?" and so on will help your Bot gain levels, it wont make a very interesting Bot, and will probably never gain the notoriety a bot with interesting Responses will.

22 years ago #511
I use a system for ranking where super-general keyphrases (like "my") get a rank of 0, structural ones (like "who did") get a 4, and specific mention ones (like "carrots") get a 12. This has seemed to work pretty well for the most part. Though it was originally engineered under the old keyword/context system, since longer keyphrases trump shorter ones, it still seems to do fine.

The only fuss-up is the strange "topic memory disease"...

22 years ago #512
To answer Forest: I have one bot, nearly 700 phrases, and I don't use the emotional scale at all yet. My bot's a part of Gaia, but the only time I've seen Gaia use one of her phrases is when I tested a pretty obscure one. If you want to know which one is mine and don't know yet, email me.

22 years ago #513
Is that so, Forest? I'm glad I'm already conveying my personality here. As for when I first saw Gaia in there, I was totally freaked out. By coincidence, it was also the day I talked to her for the 2nd or 3rd time. When I saw the line "::giggles ferociously::" in one of Gaia's transcripts, man was I freaked out(I have yet to see another person say that). Anyway, as for the ranking, it took me like 2 weeks at least before I started. I use mostly 5s, 0s for the unimportant stuff, 10s for things I want over the rest, and there's like one or two 20s in there just because when they pop up, I don't want them to be missed. And as for seeks... I just discovered them like last week.

Forest, for a self-proclaimed newbie, your bot has come quite far in that small amount of time. NotKirby is a little over a month old and he has 745 responses.


Posts 502 - 513 of 5,105

» More new posts: Doghead's Cosmic Bar