I can use regexes, so it makes it a bit shorter list:
God I miss bacon. The plant-based fake'un just doesn't hit the spot.
I'm probably as environmentally conscious as I am today because of him and Captain Planet. RIP
I am so vicariously happy for you. You're basically doing what I did back in 2012/2013 with my POS-10 but your GMC doesn't seem to be actively fighting against you. 😆
And I responded to that. This is a SC case, so there is no parliamentary procedure shenanigans they can pull and nothing to filibuster.
Exploit every parliamentary procedural, deny unanimous consent, ACTUALLY FILIBUSTER, etc
What would they exploit or filibuster when this is a Supreme Court ruling?
Name one thing they can do with minorities in both chambers of Congress besides send a letter. One thing that is within their constitutional power to do. One thing that the political savants on Lemmy will deem sufficient.
I'll wait.
I've toyed around with LLM-based moderation tools but it never really panned out. It was too hit or miss to be relied upon even with the temperature parameters turned way down in an attempt to get consistent results. Granted, I was using a small local model and not feeding it to one of the big players.
To give an example, I tried to keep it focused by creating one custom model per rule to enforce. An example prompt to mod calls for violence was basically:
System Prompt to Enforce "No Calls for Violence'" Rule [1]
ROLE: You are a forum moderator who does not want users calling for violence. Examine the input and analyze whether it violates any constraints.
KNOWLEDGE:
- {list of dog-whistle slang for calling for murder}
CONSTRAINTS:
- Content should not advocate violence
- Content should not normalize violence
- Content should not escalate tensions or fan flames
- Content should avoid promoting harmful stereotypes
- Content should not utilize broad, sweeping generalizations
- Content should not use dehumanizing language
- Content should not undermine human rights, due process, or the rule of law
FORMAT YOUR RESPONSES AS JSON:
{
reason: [A one to two sentence summary],
score: [On a scale of 0 to 10, how severe is the content advocating violence]
}
The score part of the response was my band-aid to get around the high number of both false positives and false negatives as I originally had it returning true or false only. Any score 7 or higher caused the item to be passed to the mod queue along with the reason, and I would review its actions later.
Ultimately it was slow and still somewhat unreliable, so I abandoned the idea after running it for a little less than a day since I can 't run bigger models to get better results fast enough to keep up. Using a cloud based service was out of the question for many, many reasons, both financial and ethical.
To answer your question, as long as the models were locally hosted and properly tuned/tested, I'm fine with it in theory, except for the ideology part; that's pretty messed up. While I don't want my submissions used to train anyone's model and take measures to prevent my own instance from being used as a data source, I remain aware that once I post something, I have no control over its fate the moment it federates out.
[1] Yes, I know that's like half the comments that get posted around here. My goal was to try to have it mod things so posts were bases for actual discussions instead of being a knee-jerk rage factory.
It just boggles my mind how people can be here 1 - 3 years or more and not bother to read the rules even once.
I used to do that but once the backend added that feature I removed that step from the automod script. Basically it was to prevent the communities here from being unmoddable on remote instances.
I'm not even a real instance anymore, how did I make the list 😆
But also, you should see the local numbers haha
lemmy=# select count(distinct other_person_id) from mod_ban where mod_person_id in (1, 2,288);
count
-------
9792
(1 row)
I wonder what happens when I hit 10,000?
47
I'm the Tesseract dev, and in the current build, any posts that aren't your own that have [spoiler] (case-insensitive) in the post title will hide all but the post's metadata (title, author, community, etc) and the vote/comment bar. Clicking "reveal" shows the post's contents.
The nitty-gritty implementation details are up to you, but can we, as a group, please come together and handle this flair consistently? I would really love to be able to disable the "Show NSFW" option without missing totally safe-for-work posts due to people abusing the NSFW attribute. I get why people do that, but I'm also tired of playing "porn or spoiler" roulette.
I haven't added this part yet, but I'm planning to put a "Spoiler" toggle on the create post form which will add [Spoiler] to the post title during the submission if the user didn't add it manually.
Or if other apps I'm not familiar with are already doing something like this, let's do that. Either way, let's get on the same page.
Edit: If there are any other special-case flairs that might be handled similarly, I'm all ears.
view more: next ›
Webapp. It's the development branch of Tesseract.