“What are human values, and how do we align to them?”
Very excited to release our new paper on values alignment, co-authored with
@ryan_t_lowe
and funded by
@openai
.
📝:
We were awarded a grant by
@OpenAI
, to democratize LLMs and make them wiser!
Our project, Democratic Fine-Tuning (DFT), is an alternative to Constitutional AI and the like. It gathers moral information from diverse populations to shape LLM behavior.
The youtube channel Mindful Machines has just released an *amazing* video covering our work with the Moral Graph, our vision for Meaning Economies, and why is that important for AI and the future.
Thanks so much for the great work,
@wchadly
!
(link below)
Why does so much of the world seem so dull, so full of bullshit, so lacking of life?
Why do stories of doom plague our world?
Why do most visions for the future seem to have left beauty and humanity behind?
What new stories will bring forth a more beautiful world?
I didn't realize that having no affiliation with a major institution means that your papers get stuck in arXiv purgatory 😅
in any case, our paper is now on arXiv!
...at human flourishing than markets and corporations alone, thereby replacing consumption and attention economies with economies of meaning.
We believe meaning economies have the potential to align markets with what really matters, and revitalize our ideal of progress.
@Klingefjord
's
#TAIS2024
talk on aligning
#AI
with human values begins. He'll share insights from using large language models to elicit and reconcile values across 500 Americans on divisive ethical issues. Watch live now:
#AIsafety
More news and updates soon. For now, if you want to be the first to test our prototypes, get access to our upcoming white paper and video series on ‘full stack alignment’, Wise AI demos, and other updates, join our substack above, and our discord here:
This time of year, many reflect on their values. What are yours? How can you weave your life around them?
Our work is based on a method to uncover and specify values, now packaged up in a GPT. The GPT is called “Values Discovery”.
Joe’s been using it a lot this week.
The project is led by
@edelwax
@klingefjord
as part of OpenAI's democratic process grant.
You can help us test it in a week. We'll invite people on our substack.
DFT is the first of three initiatives of our new organization, the Institute for Meaning Alignment (IMA), which exists to bring about a future where AGI and humanity cooperate on human and planetary flourishing.
Our second initiative, Wise AI, is a new architectural vision for building artificial general intelligence that's as wise is as it is smart: that understands itself in situations of moral weight, and that can develop and elaborate its values in collaboration with human groups.
The Meaning Economy Research Consortium (MERC) is our third initiative, and our most ambitious bet in socio-technical alignment.
Here, we’re researching economic mechanisms and community structures for how AI-planned economies (or market/LLM-hybrid economies) can do better...