[neom]

Not that it's a great method but just for fun I gave a large chunk of it to an LLM to process and then asked it for the 20 most disturbing or nefarious things in the chats and it was incredibly boring. Most interesting thing I learned from the files is how many gun toting americans also drive dodge chargers.

[pc86]

I would have expected nice pickup trucks or any TRD Pro trim Toyota.

I'd be curious if the LLM's own self-censorship would prevent it from reporting truly disturbing things. Maybe add one legitimately bad thing into the middle of a chat and see if it gets reported.

[neom]

I'm fairly decent a prompt engineering I think, told it was for my art project of a creative writing class and I'd hidden 20 disturbing and nefarious things in the text (made sure to inject a fake murder into the text) - fake murder then a bunch of airsoft stuff, some psychological manipulation, and it oddly surfaced...some fb cookies, heh. 2mill tokens x3 runs

[Koffiepoeder]

Have you tried querying for specific misconducts and let the LLM focus at one at a time? E.g. Find whether murders were planned or carried out, can you find any signs or plans of bomb-making, can you list all messages related to fire and arson, were any mass manipulation campaigns planned, etc, ...

I have the feeling that would probably be more effective, not sure though.

[neom]

Well my first query was "is there anything bad in here" and it basically said "no, it's a bunch of weirdos talking about guns, conspiracy theories and politics but there is nothing truly bad in there" - and then I went through a bunch of prompting for a while and very quickly got bored because at least what I was looking at, was just a bunch of americans talking about politics and guns.