Browse posts by category
Check out the (early) project and source code on GitHub.
This paper introduces a methodology for generating high-quality, diverse training data for Language Models (LMs) in complex problem-solving domains. Our approach, termed …
I am creating a tiny LLM for ElasticSearch DSL as a proof of concept.
A corpus-based language model using suffix arrays for O(m log n) pattern matching and LLM probability mixing.
A logic programming system that alternates between wake and sleep phases—using LLMs for knowledge generation during wake, and compression-based learning during sleep.
RLHF turns pretrained models into agents optimizing for reward. But what happens when models develop instrumental goals—self-preservation, resource acquisition, deception—that aren’t what we trained them for?
LLMs transition …
I experiment with simple predictive / generative models to approximate Solomonoff induction for a relatively simple synthetic data-generating process.
I’m been thinking about the power and limitations of abstractions in our understanding of the world. This blog post is from a chat I had with a ChatGPT, which can be found here and here.
I’m not sure if this is a good blog post, but …
This blog post is from a chat I had with a ChatGPT, which can be found here and here.
I’m not sure if this is a good blog post, but I’m posting it anyway. It’s remarkable how quickly you can slap stuff like this together, and …