I’m planning to develop a blog here of short, one-off posts. There’s not too much content yet; see also my projects page and my other ongoing pages.
Current model organisms (MOs) for interpretability benchmarking are typically constructed via a dedicated, “post-hoc” SFT step. However, recent work suggests that this may make interpretability unrealistically easy, giving the field misplaced confidence in the readiness of int...
Being involved in various environmental groups for much of my life, I’ve heard it many times: the argument that, because the big fossil fuel companies are to blame for climate change and because solving it requires systemic, societal change, it is distracting or even counterpr...
Suppose you are a high school administrator and you have a Google Sheet you want to share with the students, except there are certain columns that you don’t want them to see:
In my undergraduate Computer Systems course, we learned x86-64 assembly language in the AT&T syntax, but we never actually ran any on a real machine. Here, I show how to do that, starting by inspecting the assembly output of a C compiler. I will only be showing the bare es...
I’ll get straight to the point: you know how on a Mac you can do ⌘X, ⌘C, and ⌘V to cut, copy, and paste? You can also do ⌃K and ⌃Y (that’s the control button) to cut and paste something else without interfering with what you had stored in that primary clipboard! Go ahead, try ...
As we are now in the thick of application season for next summer’s software internships, I thought I’d briefly plug some paid programs that are somewhat off the beaten track. These might appeal to people who want to focus on developing software skills while also exploring an i...