The Value of Veridical Information
IntroductionI’d like to share information that is worth sharing. So I ask the question—what is valuable information?With the ubiquity of search engines and Wikipedia, much factual…
Latest
How inevitable are most accessible hard-tech startups?
(For context, I’m an undergraduate considering entering the hard-tech startup space. One concern I have is whether some of these startups are highly inevitable, and therefore…
What is Good? Antiruin and Nonabsolutism
When talking about definitions of good, it may be one’s first instinct to reach for philosophy.But there is a fairly obvious conception of good which is…
AI Mistake Seeding
I wonder if AI is being trained to make easy-to-correct mistakes so it can fix them later. That is, it ends up trained to correct its…
In Partial, Pugnacious Defense of Functional Decision Theory
I wrote a (partial) defense of FDT responding to Bentham's Bulldug. What it lacks in philosophical rigor or clear thinking it makes up for cheap jokes…
How to read tableaux, a formal system for modal logic with Kripke models
Recently I’ve been thinking a lot about a certain model of a rational agent: a proof-based agent which is triggered to act when it finds certain…
Consistency Training while Mitigating Obfuscation via Rate Matching
Sohaib Imran, Prakhar Gupta, Jannes Elstner, David Demitri AfricaLinks: Paper | Code TL;DR. Models condition their behavior on extraneous input features in undesirable ways — for…
Discovering Concept-Editing Algorithms With LLM Agents
Concept erasure is a technique that removes unwanted information from a model’s activations, but current erasure methods struggle to fully remove target concepts. In this study,…
Model access for third-parties — it's a big deal!
Over time, there might be an increasingly large gap between insider model access and outsider model access. By insiders, I mean employees at the frontier lab.[1]…
Most Current Model Organisms Leaks: Perplexity Differencing Often Reveals Finetuning Objectives
Authors: Mohammad Abu Baker, Luca Baroni, Daniel WilhelmPaper: https://arxiv.org/abs/2605.00994Code: https://github.com/z3research/ppldiff-paperTwitter thread: https://x.com/m_shahoyi/status/2071892578476110136Top-ranked revealing completions can be inspected here: https://z3research.org/This post summarizes the paper