Wouldn’t an LLM be smart enough to spot a tarpit?
LLM's don't learn on the job, they're expected to be fully-formed after completing their training. It's just too expensive for a business to invest in upgrading their workers.
Nepenthes is a tarpit to catch AI web crawlers
261–270 of 290 posts
Re: Nepenthes is a tarpit to catch AI web crawlers
#262Earlier quoted context omitted.
https://pressbooks.openedmb.ca/wordandsentencestructures/cha... I believe what the LLM replies with is in fact correct. From the standpoint of a programmer or any other category of people that are attuned to some kind of formal rigor? Absolutely not. But for any other kind of user who is more interested in the first two concepts instead, this is the thing to do.
No, I am quite sure that if you asked a random person on the street how many words are in “We the people”, they would say three.
Indeed, but consider this situation: You have a collection of documents and want to extract the first n words because you're interested in the semantic content of the beginning of each doc. You use a LLM because why not. The LLM processes the documents, and every now and then it returns a slightly longer or shorter list of words because it better captures the semantic content. I'd argue the LLM is in fact doing exactly the right thing.
Let me hammer that nail deeper: your boss asks you to establish the first words of each document because he needs this info in order to run a marketing campaign. If you get back to him with a google sheet document where the cells read like "We the" or "It is", he'll probably exclaim "this wasn't what I was asking for, obviously I need the first few words with actual semantic content, not glue words. And you may rail against your boss internally.
Now imagine you're consulting with a client prior to developing a digital platform to run marketing campaigns. If you take his words literally, he will certainly be disappointed by the result and arguing about the strict formal definition of "2 words" won't make him deviate from what he has to say.
LLMs have to navigate through pragmatics too because we make abundant use of it.
Re: Nepenthes is a tarpit to catch AI web crawlers
#263We had our non-profit website drained out of bandwidth and site closed temporarily (!!) from our hosting deal because of Amazon bot aggressively crawling like ?page=21454 ... etc. Gladly Siteground restored our site without any repercussions as it was not our fault. Added Amazon bot into robots.txt after that one. Don't like how things are right now. Is a tarpit the solution? Or better laws? Would they stop the chine…
I want better laws. The boot operator should have to pay you damages for taking down your site.
If acting like inconsiderate tools starts costing money, they may stop.
Re: Nepenthes is a tarpit to catch AI web crawlers
#264Earlier quoted context omitted.
You called the parent unintelligent yet need an LLM to show you how to run curl in a loop. Yikes.
Your assumption that I couldn't have written this myself or that I didn't make corrections to it is telling. I've only been doing dev for 30+ years lol LLMs are an accelerant, like all previous tools... Not a replacement, although it seems most people still need to figure that out for themselves while I already have
Sure, but in this case it's like driving your car 10 feet to your mailbox and then bragging about how it's an accelerant (in other words, the task wasn't remotely difficult to begin with and doesn't really warrant "accelerating"). I assume in this case your note about how it was written with an LLM was more just to spite the anti-LLM sentiment above though, which would make more sense.
Re: Nepenthes is a tarpit to catch AI web crawlers
#265> ANY SITE THIS SOFTWARE IS APPLIED TO WILL LIKELY DISAPPEAR FROM ALL SEARCH RESULTS Bug, or feature, this? Could be a way to keep your site public yet unfindable.
You can already do this with a robots.txt file
> If other pages point to your page with descriptive text, Google could still index the URL without visiting the page. If you want to block your page from search results, use another method such as password protection or noindex.
> If your web page is blocked with a robots.txt file, its URL can still appear in search results, but the search result will not have a description.
https://developers.google.com/search/docs/crawling-indexing/...
So, a robots.txt will not keep your site off of google, it just prevents it from getting crawled. (But, to be fair, this tool probably does not do this as well)
Re: Nepenthes is a tarpit to catch AI web crawlers
#266Earlier quoted context omitted.
Make it reflective, reflect it back onto an OpenAI API route.
Lol but actually this is a good way to escalate priority. Better yet, point it at various Microsoft sites that aren't provisioned to handle the traffic and let them internally escalate.
In my experience, that'd turn into a list of exceptions, rather than actually fixing the problem.
Re: Nepenthes is a tarpit to catch AI web crawlers
#267Earlier quoted context omitted.
Yeah, even beyond common decency, there's pretty strong incentives to fix it, as it's a fantastic way of having your bot's fingerprint end up on Cloudflare's shitlist.
Kinda disappointed by cloudflare - it feels they have quite basic logic only. Why would anomaly detection not capture these large payloads? There was a zip-bomb like attack a year ago where you could send one gigabyte of the letter "A" compressed into very small filesize with brotli via cloudflare to backend servers, basically something like the old HTTP Transfer-Encoding (which has been discontinued). Attacker --1kb…
> But maybe it's misconfiguration on side of cloudflare user because I can remember they at least had a WAF product in the past
They still have a WAF product, though I don't think anything in the standard managed ruleset will fire just on quotes, the SQLi and XSS checks are a bit more sophisticated than that.
From personal experience, they will fire a lot if someone uses a WAF-protected CMS to write a post about SQL.
Re: Nepenthes is a tarpit to catch AI web crawlers
#268Could a human detect that this site is a tarpit?
If so, then an AI crawler almost certainly can as well.
Re: Nepenthes is a tarpit to catch AI web crawlers
#269It feels like a Markov chain isn't adversarial enough. Maybe you can use an open-weights model, assuming that all LLMs converge on similar representations, and use beam-search with inverted probability and repetition penalty or just GPT-2/LLaMA outwith with amplified activations to try and bork the projection matrices, return write pages and pages of phonetically faux English text to affect how the BPE tokenizer gets…
What does “borking the projection matrices” and affecting the BPE tokeniser mean/look like here?
Are we just trying to produce content that will pass as human-like (therefore get stripped out by coarse filtering) but has zero or negative informational utility to the model? That would mean, theoretically if enough is trained on it would actively worsen the model performance right?
Re: Nepenthes is a tarpit to catch AI web crawlers
#270Earlier quoted context omitted.
Your assumption that I couldn't have written this myself or that I didn't make corrections to it is telling. I've only been doing dev for 30+ years lol LLMs are an accelerant, like all previous tools... Not a replacement, although it seems most people still need to figure that out for themselves while I already have
Sure, but in this case it's like driving your car 10 feet to your mailbox and then bragging about how it's an accelerant (in other words, the task wasn't remotely difficult to begin with and doesn't really warrant "accelerating"). I assume in this case your note about how it was written with an LLM was more just to spite the anti-LLM sentiment above though, which would make more sense.
That's exactly what it was meant to do. You're right, this is a trivial use case.