Lokesh L.K.S

Tech Blog

Technology, programming, and software development

J-space — reading a model mid-computation

Published on August 13, 2026 at 06:43 PM

A small, verbalizable subset of everything a model is computing — and the one fixed matrix that lets you read it out mid-layer, before the model ever finishes its forward pass. Field notes on the Jacobian lens, worked by hand.

Read more →
Mechanistic Interpretability

Transformer Architecture

Published on May 15, 2026 at 06:32 PM

It was interesting and beautiful. From input strings to numbers by a process called embedding , before embedding the prompt is tokenized. during embedding it embeds the string as well as its...

Read more →
Mechanistic Interpretability

Mechanistic Interpretability - Day 1

Published on May 13, 2026 at 07:16 PM

I did complete the prerequisite course for my day 1, I'm fully commited to it. So I'm just putting my learner's log here. I learnt a lot of einops, and really liked playing around with, It was so...

Read more →