iiinova — where your digital self keeps evolving

Nora Belrose @norabelrose.bsky.social · 24/10/2025

why don't more people become zoroastrian? it's where judaism and christianity got the idea of ethical monotheism, afterlife, and final judgment but without any of their baggage (no eternal hell, no historically questionable dogmas, etc.)

030

Nora Belrose @norabelrose.bsky.social · 11/10/2025

If we care only about appearances, outcomes, and results then AI will outcompete humans at everything If we care about the process used to create things then humans can still have jobs and meaningful lives The idea that ends can be detached from means is the root of many evils

040

Nora Belrose @norabelrose.bsky.social · 13/06/2025

if the laws of physics are fundamentally probabilistic, as they seem to be, that makes it easier to see how they can smoothly change over time

020

Nora Belrose @norabelrose.bsky.social · 12/06/2025

data attribution is a special case of data causality: estimating the causal effect of either learning or unlearning one datapoint (or set of datapoints) on the neural network's behavior on other datapoints

030

Nora Belrose @norabelrose.bsky.social · 27/03/2025

Neural networks don't have organs. They aren't made of fixed mechanisms. They have flows of information and intensities of neural activity. They can't be organized into a set of parts with fixed functions. In the words of Gilles Deleuze, they're bodies without organs (BwO).

160

Nora Belrose @norabelrose.bsky.social · 24/02/2025

Also chapter 10 where he discards the notion of the Soul but maintains the distinction between mind and brain

000

Nora Belrose @norabelrose.bsky.social · 24/02/2025

William James did a lot of good philosophy of mind in chapters 1, 5, and 6 ofThe Principles of Psychology, we've barely made any progress in 135 years 😂

030

Nora Belrose @norabelrose.bsky.social · 06/02/2025

Pro tip: if you want to implement TopK SAEs efficiently, and don't want to deal with Triton, just use this function for the decoder, it's much faster than the naive dense matmul implementation pytorch.org/docs/stable/generated/t…

080

Nora Belrose @norabelrose.bsky.social · 03/02/2025

What are the chances you'd get a fully functional language model by randomly guessing the weights? We crunched the numbers and here's the answer:

2120

Nora Belrose @norabelrose.bsky.social · 02/02/2025

we have seven (!) papers lined up for release next week you know you're on a roll when arxiv throttles you

040

Nora Belrose @norabelrose.bsky.social · 24/01/2025

deepseek now largely replacing chatgpt for me

060

Nora Belrose @norabelrose.bsky.social · 29/12/2024

Evolutionary biology can learn things from machine learning. Natural selection alone doesn't explain "train-test" or "sim-to-real" generalization, which clearly happens. At every level of organization, life can zero-shot adapt to novel situations. www.youtube.com/watch?v=jJ9O5H2AlWg

290

Nora Belrose @norabelrose.bsky.social · 28/12/2024

Truth is relative, when it comes to the physical state of the universe. But we should accept the existence of perspective-neutral facts about how perspectives relate to one another, to avoid vicious skeptical paradoxes. arxiv.org/abs/2410.13819

060

Nora Belrose @norabelrose.bsky.social · 28/12/2024

Neural networks are polycomputers in @drmichaellevin.bsky.social's sense. Depending on your perspective, you can interpret them as performing many different computations on different types of features. No perspective is uniquely correct. arxiv.org/abs/2212.10675

arxiv.org

There's Plenty of Room Right Here: Biological Systems as Evolved, Overloaded, Multi-scale Machines

The applicability of computational models to the biological world is an active topic of debate. We argue that a useful path forward results from abandoning hard boundaries between categories and adopt...

0100

Nora Belrose @norabelrose.bsky.social · 20/12/2024

If OpenAI's new o3 model is "successfully aligned," then it could probably be trusted to supervise more powerful models, allowing us to bootstrap to benevolent superintelligence.

021

Nora Belrose @norabelrose.bsky.social · 20/12/2024

Interesting to see @philipgoff.bsky.social go back and forth on the fine-tuning argument. I think the multiverse definitely can't explain fine-tuning, but it's also unclear we need an explanation at all. And God may be a more "complex" hypothesis than the physical constants themselves.

philipgoff.substack.com

My Week Without Cosmic Hope

(Photo by Tom Pumford on Unsplash)

110

Nora Belrose @norabelrose.bsky.social · 11/12/2024

How do a neural network's final parameters depend on its initial ones? In this new paper, we answer this question by analyzing the training Jacobian, the matrix of derivatives of the final parameters with respect to the initial parameters. arxiv.org/abs/2412.07003

49219

Nora Belrose @norabelrose.bsky.social · 10/12/2024

Bombshell new paper on the simulation argument, multiverses, and cosmological fine-tuning: "...self-locating credences are ‘subjective’ in the sense that they are not rationally constrained by anything at all, except possibly the requirement of probabilistic consistency." arxiv.org/abs/2409.05259

arxiv.org

Against Self-Location

I distinguish between pure self-locating credences and superficially self-locating credences, and argue that there is never any rationally compelling way to assign pure self-locating credences. I firs...

040