Dec 10 in Boulder: Local Weights & Measures
On December 10 I'm hosting the AI/ML Engineering meetup in Boulder: three mini demos that open up model weights, from LLM mind reading to LoRA remixing.
On Thursday, December 10, I’m hosting the AI/ML Engineering meetup (an RMAIIG subgroup) in Boulder: Local Weights & Measures: LLM Mind Reading, Brain Surgery, and LoRA Remix.
- When: Thursday, December 10, 9:00 to 11:00 AM MST
- Where: Founder Central by Sweater, 2000 Central Ave #100, Boulder, CO 80301
- RSVP: on Meetup
Usually we treat model weights as a black box: billions of numbers we download and never look at. This month we open the box with three mini demos.
LLM Mind Reading. A recent MIT paper had LLMs work out what small neural networks were trained to do just by reading their raw weights, and Simon Willison replicated it with a local model on his laptop. How does that work, and what are the limits?
LLM Brain Surgery. In the same paper, GPT-6 Astra found what was sending a small Qwen model into reasoning loops, and fixed it by editing 0.016% of its weights. We’ll look at what was actually changed, and how to check for edits like that yourself.
LLM LoRA Remix (time permitting). LoRAs as a mixing board. We’ll mix, subtract and turn up adapters live on image and text models.