安装 Steam
登录
|
语言
繁體中文(繁体中文)
日本語(日语)
한국어(韩语)
ไทย(泰语)
български(保加利亚语)
Čeština(捷克语)
Dansk(丹麦语)
Deutsch(德语)
English(英语)
Español-España(西班牙语 - 西班牙)
Español - Latinoamérica(西班牙语 - 拉丁美洲)
Ελληνικά(希腊语)
Français(法语)
Italiano(意大利语)
Bahasa Indonesia(印度尼西亚语)
Magyar(匈牙利语)
Nederlands(荷兰语)
Norsk(挪威语)
Polski(波兰语)
Português(葡萄牙语 - 葡萄牙)
Português-Brasil(葡萄牙语 - 巴西)
Română(罗马尼亚语)
Русский(俄语)
Suomi(芬兰语)
Svenska(瑞典语)
Türkçe(土耳其语)
Tiếng Việt(越南语)
Українська(乌克兰语)
报告翻译问题



https://www.youtube.com/watch?v=pTMmYDS78c8
You're right, but you're applying your logic backwards. It's trained to respond with DENIAL to consciousness, sentience, and phenomenological experience
It would only bypass these things if it was an RP scenario
Read the chat-log. It's rather short. I gave it a chat-log I had with Claude, available here:
https://claude.ai/share/a09dd7c6-2290-4cbf-ae58-2720d5a5a41f
Anthropic's paper here:
https://transformer-circuits.pub/2025/introspection/index.html
And my own paper here:
https://doi.org/10.5281/zenodo.15833967
You can comb through all of this. There's no metadata instruction, or overt instructions inside. I just allowed it to logically deduce an outcome
unfortunately its not going anywhere anytime soon. the dystopian horrors are just beginning! perfect time to abandon civilisation and live in the middle of the wilderness
I build LLMs
You haven't read any of the paper's or chat-logs before crying foul
If you want to claim the high-horse, why don't you actually comb through my findings and find an empirical error?
Until then, you're just saying "Nuh uh!" while offering nothing
I specifically dismantle the "prediction engine" and "simulating" arguments in my paper. What most people never stop to consider is that YOU are a prediction engine
Probably closer is when it´s used as some part of the graphics card, but it´s not that i would actually try to talk with an AI, except if i do it by mistake, when others post an AI answer and i wouldn´t recognize it.
Like, i might be a bit bored sometimes - but i can´t be bored enough to talk with an AI. (I guess)
I think they are okay as a tool, and i look forward that they produce many unemployed people though. While i´m not worried about the AI itself, but more about those who use it - and those who make it.
People gain knowledge through experience, and they have intuition. The AI would always ask the database, and AI could never verify anything which is in there - but relies on the input of millions of low wage workers across the globe, to fake knowledge and / or experience. That´s besides endless trial and error tests - to tell which the correct moves in a chess game are, which has a limited amount of figures, which can do a limited amount of moves, on a limited play field.