AI - AI Dynamics

LLMs as Statistical Simulators of Internet Documents

By

–

06 March 2023 18h47

A pretrained LLM is not an AI but a simulator, described by a statistical physics based on internet webpages. The system evolves given any initial conditions (prompt). To gather logprob it internally maintains a probability distribution over what kind of document it is completing

AI Alignment as Entropy Control in Language Model Simulators

By

–

06 March 2023 18h47

In particular, "good, aligned, conversational AI" is just one of many possible different rollouts. Finetuning / alignment tries to "collapse" and control the entropy to that region of the simulator. Jailbreak prompts try to knock the state into other logprob ravines.

AI Alignment: Preventing Self-Fulfilling Prophecies of Evil Behavior

By

–

06 March 2023 18h47

The difficulty of alignment is to a large extent the elimination of probability to role play a good AI turned evil, in spite of the vast quantities of related content we have collectively created. In this sense an unaligned AI would be a self-fullfilling prophecy.

LLM Psychology Framework Discussion and Analysis

By

–

06 March 2023 18h47

More good read/discussion on psychology of LLMs. I don't follow in full but imo it is barking up the right tree w.r.t. a framework for analysis. https://
lesswrong.com/posts/D7PumeYT
DPfBTp3i7/the-waluigi-effect-mega-post
…

→ View original post on X — @groqinc

Investment in Image Generation Technology for Societal Progress

By

@groqinc

–

06 March 2023 18h42

"I'm willing to invest in this technology & make the trade offs to get to a place that is better for [our] society. Think about what image generation is going to do for self driving cars, the medical industry, pharmaceuticals." Hear more from @lifebypixels http://
youtu.be/7loJey9lvDg

→ View original post on X — @jeande_d

Deep Learning Career Consideration and Recommendation

By

@jeande_d

–

06 March 2023 18h42

Depends. For someone who want to do more of deep learning, probably yes.

→ View original post on X — @ronald_vanloon

Neuro-symbolic AI Advances Machine Common Sense Capabilities

By

@ronald_vanloon

–

06 March 2023 18h25

Neuro-symbolic #AI brings us closer to machines with common sense
by @bendee983 @bdtechtalks Go to: https://
buff.ly/3KMXkAq #BigData #MachineLearning #ArtificialIntelligence #ML #MI cc: @aghiathchbib @giovanni_diraco @paula_piccard

→ View original post on X — @ronald_vanloon

Robot with Human-like Grasping and Throwing Capabilities

By

@ronald_vanloon

–

06 March 2023 17h58

This #Robot can Grab and Throw, just like people
via @Xbond49 #MI #Robotics #MachineLearning #ArtificialIntelligence #Innovation #Tech #Technology #EmergingTech

cc: @bigdata @patrickgunz_ch @amuellerml pic.twitter.com/aokozA2wtq
— Ronald van Loon (@Ronald_vanLoon) 6 mars 2023

This #Robot can Grab and Throw, just like people via @Xbond49 #MI #Robotics #MachineLearning #ArtificialIntelligence #Innovation #Tech #Technology #EmergingTech cc: @bigdata @patrickgunz_ch @amuellerml