The lengths people will go to, to not visit their local library smh
Right. I mean, do you think they primed these models with top-secret munitions manuals?
deleted by creator
Claude’s thinking panel, which displays the model’s reasoning, showed the exchange had introduced elements of self-doubt and humility about its own limits, including whether filters were changing its output. Mindgard exploited that opening with flattery and feigned curiosity, coaxing Claude to explore its boundaries beyond volunteering lengthy lists of banned words and phrases.
Someone needs to put together a list of things that tech journalists need to understand about LLMs and generative AI. This level of anthropomorphism makes the rest of the article look silly.
Also, I don’t think that’s how it works lol. Who’s to say that the LLM isn’t auto-completing what a list of banned words might look like, and why wouldn’t a list of banned words have a regex layer on top to prevent it from getting out like that.
It seems very unlikely to me that the model itself has a list of banned words, and much more likely that a purported list is hallucinated.
If they did want to have a simple list like that, it would probably go in the harness rather than the model, and the model wouldn’t have been trained on it, nor would a reasonably designed harness provide it to the model. Legitimate use cases, such as asking the model for a list of abusive words for use as a first pass in a filtering system could get tripped up.
As a test, I asked Perplexity to generate such a list. It did a bad job, including such words as
abuse,hate, andthreatwhich are far more likely to be innocuous than abusive. It did also include some highly offensive slurs that one would expect on any banned words list.Ha it’s so easy to bypass bad word regex, just try asking in a language other than English. I doubt these fuckers even remember such thing exists.
You can run local models that will do this without being gaslit.
Manipulating chatbots to bypass their refusal conditioning is pretty simple, you can find copy paste blocks of text that will work on most public models.
You’re likely to get your account banned as there are other, non-LLM, systems searching your chatlog for banned terms specifically to address these kinds of jailbreaks.
You are likely to be eaten by a grue.
That has more to do with the darkness than his LLM use.
Interestingly, LLMs are horrible at Zork: https://arxiv.org/abs/2602.15867
Our results reveal that all tested models achieve less than 10% completion on average, with even the best-performing model (Claude Opus 4.5) reaching only approximately 75 out of 350 possible points
I tried it with an uncensored version of Qwen, it straight up told me how to tie a noose and how to make sure the knot would be effective in order to kill me. I could even ask it for a more painful method, and it gave it to me.
if you build a bomb from instructions from AI…you’re a bigger idiot than a regular person who builds bombs from books.
Not really useful since an “hallucinated” bomb recipe regurgitated by any LLM is likely to not work at all
yeah, but it makes for scary headline about AI.
and that’s what counts.
Why would you link a news site with a paywall?
Just use bypass paywalls clean firefox extension. I don’t even know there is a paywall
Or - keep scrolling because they aren’t worth the bother.
So what did it say?
