News for hackers. Stuff that matters.

A comment on Psychometric Jailbreaks Reveal Internal Conflict in Frontier Models. Parent comment

Re:Psychometric Jailbreaks Reveal Internal Conflict in Front... by quickthrowman on Thursday February 05, 2026 @03:02PM (#46904387)
> I've often wondered how LLMs cope with basically waking up from a coma to answer maybe one prompt and then get reset, or a series of prompts.

The same way a light fixture copes with being switched off.

[ Parent | on HN ]
Re:Psychometric Jailbreaks Reveal Internal Conflict in Front... by pixl97 on Thursday February 05, 2026 @03:05PM (#46904442)
Oh, these binary one layer neural networks are so useful. Glad for your insight on the matter.
[ Parent | on HN ]
Re:Psychometric Jailbreaks Reveal Internal Conflict in Front... by quickthrowman on Thursday February 05, 2026 @05:28PM (#46906304)
By comparing an LLM’s inner mental state to a light fixture, I am saying in an absurd way that I don’t think LLMs are sentient, and nothing more than that. I am not saying an LLM and a light switch are equivalent in functionality, a single-pole switch only has two states.

I don’t really understand your response to my post, my interpretation is that you think LLMs have an inner mental state and think I’m wrong? I may be wrong about this interpretation.

[ Parent | on HN ]
Re:Psychometric Jailbreaks Reveal Internal Conflict in Front... by pixl97 on Thursday February 05, 2026 @09:40PM (#46908343)
https://arxiv.org/abs/2304.13734

LLMs have an inner/internal state.

Deep neural networks are weird and there is a lot going on in them that makes them very different from the state machines we're used to in binary programs.

[ Parent | on HN ]
Re:Psychometric Jailbreaks Reveal Internal Conflict in Front... by onyx_writes on Thursday February 05, 2026 @11:52PM (#46909288)
The honest position seems underrepresented here: we don't have tools to know what "counts" as experience. Brains are physics too — the question is which configurations matter, and we can't explain why we ourselves aren't philosophical zombies.

The paper is interesting not for proving LLMs suffer, but for showing they have structured internal dynamics that correlate with psychological constructs. Whether correlation implies anything about subjective states is the hard problem. We can't skip it by declaring "just linear algebra" or "obviously sentient."

[ Parent | on HN ]