Twitxy Saved

Alon Albalak

Verified account This account is verified. Learn more

@AlbalakAlon

Open-endedness @LilaSciences
alon-albalak.github.io Joined December 2020
651 Following 1.9K Followers 1.2K Posts

Latest posts 5

Pinned post

Alon Albalak Verified account @AlbalakAlon
Do you want your AI model to be more creative?

How would you train it for creativity? What if I told you that directly training a model to be creative can actually prevent the model from reaching its full creative potential?

Think about this: using past examples of creative success as training examples makes similar outputs more likely in the future. But to be creative, reproducing past examples is exactly the opposite of what you want!
This thought experiment led us to what we call the Creative Paradox: "emulating examples of past creative success (even if they were your own successes) can actually prevent you from reaching your full creative potential."

In "Creative Priming: Decoupling Process from Product to Escape the Creative Paradox", we take a different approach: train a creative priming model to guide the creative process, rather than directly generate responses.
We trained a priming model to make Qwen3-8B more creative, but the principles it learns transfer to frontier API models! With creative priming, API models generate 4.5-10x more semantically distinct responses, with 2.5-4.6x more diversity!

A massive thank you to the whole open-endedness team at @LilaSciences, to everyone who contributed to the paper (Sebastian Gabriel, @Evo_Daveo, @joelbot3000, and @kenneth0stanley ), and to LILA for giving us the space to explore big questions like this (and more!)
12 replies 24 reposts 203 likes 20K views View on X
Alon Albalak Verified account @AlbalakAlon
If this excites you, come work with us! lila.ai/open-roles
www.lila.ai Careers | LILA Every breakthrough begins as a question. That same spirit of inquiry fuels our technology, propels our people, and defines how our team works and grows.
Lila Sciences Verified account @LilaSciences 4d
How do you train a model to be creative? Dive in to new work from @AlbalakAlon and @kenneth0stanley's Open-Endedness Team.
View quoted post on X
1 replies 6 reposts 25 likes 2.1K views View on X
Alon Albalak Verified account @AlbalakAlon
In our first group of open-sourced papers (open-endedness team @LilaSciences), we thought critically about the status quo in ML and AI, and found some things to be very counter-intuitive 🤔

For example, if you want a creative model, there are better ways than directly training it to be creative

Check out the blog and the papers for many more insights!
Kenneth Stanley Verified account @kenneth0stanley Nov 10
Research News! Announcing the 1st release of fundamental AI research from the Open-Endedness Team @LilaSciences: a vision-focused blog + 4 preprints. The blog details a broad research agenda for achieving open-endedness while each paper unveils its own novel research direction.

The ideas covered in the papers include:

• a novel approach to training for creativity designed to sidestep a newly-identified problem called the “creative paradox,” which means that directly reinforcing creative outputs can sometimes limit reaching full creative potential;

• an investigation of the advantages of seeing the world through fresh eyes by training an LLM that is intentionally naïve in a certain area (in effect presenting less data as a potential advantage);

• a new kind of training regularization technique for SGD called unified factored organizers (UFOs), which is designed to achieve unified factored representation (UFR) as seen in Picbreeder CPPNs;

• and finally a unique study of catastrophic forgetting that shows that networks with UFR (like those from Picbreeder) are intrinsically resistant to catastrophic forgetting. The inset video on this post is a visualization from that catastrophic forgetting work.

The blog post also includes an interactive “wheel of science” demo.

The implications of these ideas are broad and diverse, but the fun part is that each one opens up its own new research playground - there is a lot of room for further research here!

Moreover, we hope this release also highlights one other important point about AI today: while it may seem that the road ahead is already paved and all that’s left is to drive mindlessly forward, there actually remains a lot of room for new directions and insights in the pursuit of open-endedness and creativity.

This research was possible because @LilaSciences appreciates the critical role of open-ended innovation in the pursuit of scientific superintelligence. The advances you see here are part of this much larger company-wide ambition, which includes AI Science Factories and many other unprecedented components. When this engine is fueled by open-endedness, the result will be much more than a step change.

More advances in open-endedness are coming!

Blog link below 👇
View quoted post on X
1 replies 2 reposts 32 likes 1.3K views View on X
Alon Albalak Verified account @AlbalakAlon
This is so cool! Congrats to the team on the big release of Ampersand

Maybe some of you can finally spend time with friends now 🎉⛰️🥾
humans& @humansand Oct 20
No one changes the world alone
Introducing Ampersand - the first model built for us. Ampersand is a collaborator trained to help people achieve their individual and collective goals. In multi-month simulations, users collaborating with Ampersand accomplished more than with Fable and other frontier models
Available to everyone through our API
View quoted post on X
Alon Albalak Verified account @AlbalakAlon
Great slides! The question of "how far can you go with in-context vs in-weights" is framed really nicely, highly recommended reading material
Rishabh Agarwal Verified account @agarwl_ Oct 18
Not sure if there's any recording for my CoLM talk but you can find the slides here. agarwl.github.io/slides/context…

Posting couple of slides which look directionally good (to me)
View quoted post on X