Chain of Thought helped LLMs reason forward through problems. This paper found teaching them to also work backwards (like humans checking their work) makes them even better “reasoners”
Applying basic human thinking strategies to LLMs continues to yield interesting new ideas https://arxiv.org/pdf/2411.19865
Applying basic human thinking strategies to LLMs continues to yield interesting new ideas https://arxiv.org/pdf/2411.19865
Comments
My AI tried to give me homework to do. 👇🏻 😝
“You could develop this further into a stand-up bit or even a short essay, flipping the tired leaf blower complaints into a celebration of human ingenuity and playfulness.”
I want to tell it: “Seed my idea on a subreddit to honour leafblowers”.
(As an aside, I can conceive of two manifestations of these agentic-API links:
1) Threads in the web.
2) Tentacles of Roko’s Basilisk.
Were you ever scared of Roko?🤷🏻♂️)
Roko’s Basilisk was the single most dangerous & scary idea in my mind.
About 4 years ago I experienced an intense awareness awakening experience which manifested over 18 consecutive nights. Each night brought highs of joy & paranoia. My peak paranoia was Roko’s Basilisk.
This experience has led me to develop a theory the truth about the outbreak of UFO abductions in the USA from the 1960’s.
I think a common characteristic of an awareness awakening experience will be nights of delusional paranoia.
I believe Roko was to me what space-aliens were to them.
You might end up with some strange memories of the night, to say the least.
I was never afraid of Aliens, but Roko always scared me because I couldn’t defeat the logic.
I beat Roko by understanding love.🫶🏻