AI referencing Literotica

[EmilyMiller]

Ironically the following is from Wikipedia:

A trauma trigger is a psychological stimulus that prompts involuntary recall of a previous traumatic experience. The stimulus itself need not be frightening or traumatic and may be only indirectly or superficially reminiscent of an earlier traumatic incident, such as a scent or a piece of clothing.[1] Triggers can be subtle, individual, and difficult for others to predict.[2][3] A trauma trigger may also be called a trauma stimulus, a trauma stressor or a trauma reminder.[4][5]

—

When used in on-line disagreements, the intent is often to belittle your opponent, or to suggest that they are reacting emotionally rather than rationally. It is often also an attempt by one party to shift blame for their own actions and statements onto their opponent, rather than taking responsibility for them. “Your reaction is the issue, not what I did / said.”

 

[ADirtyPerv]

Ironically the following is from Wikipedia:

A trauma trigger is a psychological stimulus that prompts involuntary recall of a previous traumatic experience. The stimulus itself need not be frightening or traumatic and may be only indirectly or superficially reminiscent of an earlier traumatic incident, such as a scent or a piece of clothing.[1] Triggers can be subtle, individual, and difficult for others to predict.[2][3] A trauma trigger may also be called a trauma stimulus, a trauma stressor or a trauma reminder.[4][5]

—

When used in on-line disagreements, the intent is often to belittle your opponent, or to suggest that they are reacting emotionally rather than rationally. It is often also an attempt by one party to shift blame for their own actions and statements onto their opponent, rather than taking responsibility for them. “Your reaction is the issue, not what I did / said.”

Not to kick things harder, but IMO you're reading @Publius68 wrong. Their words were:

And @EmilyMiller I do apologize for trigering your I am sure just defense of Wikipedia.

The trigger was related to your defense of wiki, not to your emotions. The rest of their response (which was self-deprecating instead of attacking) would further reinforce that they were not intending to belittle at all.

 

[EmilyMiller]

Not to kick things harder, but IMO you're reading @Publius68 wrong. Their words were:


The trigger was related to your defense of wiki, not to your emotions. The rest of their response (which was self-deprecating instead of attacking) would further reinforce that they were not intending to belittle at all.

Then don’t misuse a triggering word like trigger in the first place. I wasn’t triggered by their BS claim, but I sure was by using that word. They might as well have started with, “My dear lady, let me explain…”

 

[Actingup]

This. Wikipedia isn't perfect by any stretch, but it's not intended that users should take it on trust; it's intended that editors should provide citations and readers should check those citations. When used in that mode, it's pretty good. One still has to watch out for subtler biases and omissions, but that applies to pretty much anything out there.

Incidentally, most LLMs are trained on Wikipedia among other things. The idea that they'd be more trustworthy than their training data is an intriguing one.

There's a parallel in ensemble numerical modelling, which tends to verify better than any single model as different insights are added to the mix. For what it's worth, I've seen some outstanding Wikipedia articles, and I've also been very frustrated by wrong ones where attempts to correct the article from my direct inside knowledge of the subject material have been rebuffed because there wasn't a citable source. Of course, peer-reviewed literature also has flaws. Britannica has flaws. Truth Social has flaws. And all of these are subject to manipulation. But Wikipedia is a wonderful resource.

Also, Grok is evil.

 

[SmilingLez]

Also, Grok is evil.

100% agree

 

[EmilyMiller]

Also, Grok is evil.

Everything to do with Elon Scum is evil.

 

[old_prof]

There's a parallel in ensemble numerical modelling, which tends to verify better than any single model as different insights are added to the mix. For what it's worth, I've seen some outstanding Wikipedia articles, and I've also been very frustrated by wrong ones where attempts to correct the article from my direct inside knowledge of the subject material have been rebuffed because there wasn't a citable source. Of course, peer-reviewed literature also has flaws. Britannica has flaws. Truth Social has flaws. And all of these are subject to manipulation. But Wikipedia is a wonderful resource.

Now this is the kind of sweet seductive talk I was hoping to see in the hallucinating readers thread.

 

[Zootonius]

And it is way more reliable than Wikipedia.

Grok is trained on Twitter data.

These two points are very much in contradiction.

 

[old_prof]

I think early on in Wikipedia's existence, a lot of teachers, librarians and academics demonstrated very elitist thinking that anything that included input from the general public was untrustworthy. It was far from a universal opinion even at the time, and, in my experience, has diminished substantially as people understand what the site really does do. I think many of that crowd who still do presume its inaccuracies have substantial insecurities about their own expertise.

The second major group of people I have known to badmouth wikipedia are primarily upset because it will not support their own fringe theories, from flat earth to modified gravity to politics.

Note that I am not saying that OP is in either of these camps, just that these camps exist in substantial numbers.

Wikipedia is far from perfect as a resource, but as many have said, on the whole it's pretty good.
And at least I have never heard of it hallucinating citations, which is far too prevalent from all of the LLMs.

 

[EmilyMiller]

I think many of that crowd who still do presume its inaccuracies have substantial insecurities about their own expertise.

 

[SkyBubble]

Grok is trained on Twitter data.

Among other things. Which makes it way more relaiable than Wikipedia, as Publius68 said.

 

[EmilyMiller]

I sense a political divide here - it was nice when facts were apolitical.

 

[anthrodisiac]

I sense a political divide here - it was nice when facts were apolitical.

I think the last time facts were apolotical was when there was a single lifeform on the whole planet. Then it divided, and we've been dealing with this shit ever since.

Weirdly, their first fight was also about Wikipedia...

 

[EmilyMiller]

I think the last time facts were apolotical was when there was a single lifeform on the whole planet. Then it divided, and we've been dealing with this shit ever since.

Weirdly, their first fight was also about Wikipedia...

There appeared to be a brief period where knowledge was apolitical. Until it revealed some inconvenient truths. Now we have the opinions of the inexpert somehow being relevant. It feels rather end of days.

 

[anthrodisiac]

There appeared to be a brief period where knowledge was apolitical. Until it revealed some inconvenient truths. Now we have the opinions of the inexpert somehow being relevant. It feels rather end of days.

If it's heartening at all, what we're seeing isn't really all that different from most of human history, it's just playing out on a much larger scale. Facts are a tool of power, and thus subject to political whims, because it sets the ground truths that prop up power structures.

In American history, really the period of about 60 years from (roughly) the end of WWII to the late 90s, when most people got their news from a handful of sources (primarily the handful of TV channels), is where people agreed on the basic underlying facts. Granted, a LOT of shit was being swept under the rug, so even then, it just appeared like we all agreed on the "facts," even if they were, in fact, not true, or half-truths.

But that's just from an American lens. You consider that for every "fact" we had in the US, the USSR had their own set, other countries had their own.

That might be even more disheartening, now that I think about it...

Point is, there really hasn't been an actual period of human history where facts weren't disputed or used as political tools. It seems worse because there's a lot more visibility and transparency, and the information ecosystem is so fractured that there are massive incentives to set up silos of information to capture attention and money and power. Normally, it would've been under the iron fist of whatever king or emperor or dictator happened to be in power, they dictated what was true and what wasn't. Instead of a handful of smaller kingdoms telling their subjects what was true, we now have it on a global level, with self-sorted kingdoms where people pick and choose what they want to believe for their own benefit or to fit their worldview or whatever. Instead of power massively restricting what people can know, we have the exact opposite, where there's so much information, misinformation, and disinformation, that the set of facts has expanded from whatever a king declared to such a huge potential subset that it's impossible to sort through.

Is this better? Hard to say. Is it better to have a hundred sources telling you the sky is any number of different colors instead of a king tell you that the sky is purple and you have no choice but to nod your head, lest it be cut off?

It's for sure a problem, though. It's why teaching critical thinking is so important. Because it's hard to say what's true right off the bat, but you can teach people to investigate and think, to do their research and find reputible sources, to avoid the instinct to crouch into a defensive posture when they come across something they disagree with, and instead to be open and curious. And, most importantly, to be willing to admit they might have been wrong about something.

This has been a serious, thoughtful reply, brought to you by an archaeopteryx. You are now returned to Anthy's usual chaotic meerkat nonsense.

 

[EmilyMiller]

Is this better? Hard to say. Is it better to have a hundred sources telling you the sky is any number of different colors instead of a king tell you that the sky is purple and you have no choice but to nod your head, lest it be cut off?

The sky is objectively black right now. It’s an irrefutable observation in the U.S. at one in the morning. There is such a thing as objective truth. The problem occurs when objective truth isn’t want people want to hear.

 

[SkyBubble]

Wikipedia is OK for lists like Best Musical winners year by year. It's not much good for anything else.

 

[TheLobster]

Anyone who still thinks Wikipedia is reliable or trustworthy for anything that’s more concrete and real world-adjacent than topology of Kleene manifolds in non-Euclidean spaces (*) should really read this.

(*) Yes, that’s an exaggeration, but you know what I mean.

 

[ADirtyPerv]

Anyone who still thinks Wikipedia is reliable or trustworthy for anything that’s more concrete and real world-adjacent than topology of Kleene manifolds in non-Euclidean spaces (*) should really read this.

(*) Yes, that’s an exaggeration, but you know what I mean.

That's a pretty niche point with which to dismiss the entirety of Wikipedia. Eh. Nothing is without bias or propaganda. There's a blurb at the end addressing a response to the blog post you linked to:

"“[T]he only way to improve Wikipedia is for more people, of all stripes, to get involved. […] [T]o condemn and abandon Wikipedia is to give up on something beautiful, and push us further along a path whose destination we will all — right, left, centre, whatever — regret profoundly.”

I think that's a much more valid stance to take. Acknowledge where there is bias, check the sources, but don't throw the baby out with the bathwater.

 

[SomaSlave]

Jeeze. I just tested this myself and it's slightly concerning. First I asked it "Who is the Literotica author OddLove"

It gave my basic and publicly available information and summarized my writing style and types of stories.

Then I clicked on the suggestion "Recommend OddLove's Top Stories" and it gave me recommendations for my main series Kind and Cruel Dominatrix.

Then I clicked the suggestion "Analyze Kind and Cruel Dominatrix"

It then did exactly that. First a quick summary, then an 'Overall Plot Arc' with a 'Light Spoiler' warning.

And it went to talk about the story and characters in detail. Details that are only available IN the story. Not the comments.

Meaning it read the story.

So, either it read my story upon my request to 'Analyze' it, or it literally scrubbed all of Literotica and all our stories are already fed into Grok.

The most annoying part of this for me is, I as an author didn't give Elon permission to add my stories into his AI's database.

And dinosaurs in the government don't understand AI and billionaire tech bros along with their spineless cucks don't respect consent, so nobody can do anything about this.

Maybe the solution is to be fickle and submissive and handwave the problem away like "Well, it is what it is" and accept our stories are being used to train AI without our consent, and without compensation.

Here's what Gemini had to say about. I prompted, "Tell me what you know about Literotica author SomaSlave." This was its reply:

SomaSlave is a prolific author on Literotica, primarily known for writing within the Science Fiction, BDSM, and Mind Control genres. Their work often explores themes of dystopian futures, societal restructuring, and psychological conditioning.

It also praised my story "The Soma Slaves of New Eden." The only problems are that I've never written in the Science Fiction or Mind Control categories, and I couldn't locate that story in Literotica. Are my stories so boring that it had to hallucinate better ones?

 

[SomaSlave]

Here's what Gemini had to say about. I prompted, "Tell me what you know about Literotica author SomaSlave." This was its reply:

SomaSlave is a prolific author on Literotica, primarily known for writing within the Science Fiction, BDSM, and Mind Control genres. Their work often explores themes of dystopian futures, societal restructuring, and psychological conditioning.

It also praised my story "The Soma Slaves of New Eden." The only problems are that I've never written in the Science Fiction or Mind Control categories, and I couldn't locate that story in Literotica. Are my stories so boring that it had to hallucinate better ones?

Here's a followup. After I corrected Gemini on its hallucination, it found me and provided this trenchant analysis:

Overview of SomaSlave's Work​

SomaSlave is known for a very specific niche on Literotica, primarily within the Gay and BDSM categories. Unlike the general "mind control" tropes I initially suggested, this author’s work is deeply centered on:

• Gay Submissive/Slave Narratives: The stories typically follow a protagonist (often "Soma") through intense, transformative experiences of submission.
• The "Soma's Journey" Series: This is their central body of work. It is a long-running narrative that chronicles the character's evolution from an ordinary individual into a deeply conditioned submissive.
• Literary Ambition: Titles like "The Apotheosis of Soma" (Apotheosis meaning the elevation of someone to divine status) suggest a style that is more philosophical or "journey-focused" than standard short-form erotica. It deals with the total psychological and physical surrender of the character.

Where to Find the Stories​

On Literotica, you can find them by searching for the author Somaslave (try it as one word if the space doesn't work).

• Category: Most of these stories are filed under Gay, BDSM, or Submissions / Personal Accounts.
• Tone: The stories are often described by readers as intense and dark, focusing on "total power exchange" (TPE) and the eroticization of being "used" or reshaped by others.

Apart from missing out on another category in which I've written extensively, it provided some interesting insights into my writing. It's nice to know that I have Literary ambition. It's also not nice knowing that Gemini is also training on Literotica. Interestingly, Claude appears not to be trained there. It knows of the site, but couldn't identify any specific authors. Its guardrails are still good.

 

[TheLobster]

That's a pretty niche point with which to dismiss the entirety of Wikipedia.

If you have more time, you can go through a much longer article by the same blogger which illustrates a practical way in which systemic issues with the way Wikipedia sorts the supposed wheat of truth from the chaff of falsehoods enables abuse on a massive scale by just a single individual embedded within the organization.

Ultimately that’s the main problem with it: the fact that despite the claims that “anyone can edit it”, Wikipedia is controlled by a small handful of highly entrenched admins and moderators who allow their own ideological biases to seep in (to the point of running an apology tour for the worst dictator in history) under the seemingly innocuous umbrella of the site’s policies, most notably the infamous “reliable sources”.

Which, ironically, is the exact opposite of the early arguments against Wikipedia: that it is an unreliable Wild West because “anyone can edit it”. They can’t, not anymore, and that’s the problem.

 

[Purechicanery]

IMO it is simultaneously true that:
a. Wikipedia has a bunch of issues with accuracy because it is contributed to and curated by humans with biases
b. This is true of literally every other human created reference
c. Wikipedia is way more reliable than any LLM, and contains built in references to fact check

Anyway, my immediate reaction when someone says that they consulted an LLM to find something out, is usually something along the lines of "OK well I can safely disregard the conclusion you've come to."

 

[ADirtyPerv]

If you have more time, you can go through a much longer article by the same blogger which illustrates a practical way in which systemic issues with the way Wikipedia sorts the supposed wheat of truth from the chaff of falsehoods enables abuse on a massive scale by just a single individual embedded within the organization.

Ultimately that’s the main problem with it: the fact that despite the claims that “anyone can edit it”, Wikipedia is controlled by a small handful of highly entrenched admins and moderators who allow their own ideological biases to seep in (to the point of running an apology tour for the worst dictator in history) under the seemingly innocuous umbrella of the site’s policies, most notably the infamous “reliable sources”.

Which, ironically, is the exact opposite of the early arguments against Wikipedia: that it is an unreliable Wild West because “anyone can edit it”. They can’t, not anymore, and that’s the problem.

Thanks for a balanced response. It's the Internet, so I wasn't expecting any clarity of thought at all. Color me chuffed, or whatever the Brits say.

I'll admit that Wikipedia is obtuse in its management of data / knowledge. That kind of thing is ripe for abuse, as power corrupts. With knowledge being power, and Wikipedia being a significant source of knowledge, corruption in the control of that knowledge - however minor it may be - is significant.

But I will say that anyone taking political commentary from Wikipedia is immediately mistaken. The site can (and should) summarize the general consensus without attempts to pick one unless the consensus is profound in one way or another. I'll give you that the Mao legacy section on its own is a little more biased than I'd like, and I attribute that to the corruption inherent to any place of power.

But you can see that section as summarizing the political commentary surrounding Mao, but with a bias. All of the rest of the data in that wiki entry is pretty solid. So as long as people take the opinionated sections with a grain of salt, the rest is totally valid.

Baby, bathwater.

 

[Bramblethorn]

And at least I have never heard of it hallucinating citations, which is far too prevalent from all of the LLMs.

I have on rare occasions known people to fabricate Wiki citations. There was one guy in particular who used to insert bogus material cited to sources that could not easily be checked (print-only records in some out-of-the-way repository, that kind of thing) and which turned out not to exist, and plenty of others who cited sources that didn't actually support their claims.

That said - every medium has its bad actors, and I've seen just as egregious stuff in peer-reviewed academic journals. A much bigger problem is well-meaning folk who aren't very good at figuring out what info is reliable enough to go in, and people who get territorial about articles and resist anybody else's ideas.

There's a parallel in ensemble numerical modelling, which tends to verify better than any single model as different insights are added to the mix.

That's an interesting point, though I'm not sure how directly it translates to LLM behaviours.

AFAICT, what LLMs learn is weighted by how many times it sees the same patterns in the training data (as well as any intentional weighting) and there are a lot of Wikipedia mirrors, clones and translations out there, so I suspect the average LLM is still going to be very heavily influenced by Wiki content. If you're running an ensemble but half the ensemble is clones of a single model...