ada1981

ada1981

14 dagen geledendiscuss

Sorry to disapoint!

I suppose it's relative.

It was essentially 2 days of work using AI to help us get it all together and we were able to release it within 48 hours of the research dropping. It wouldn't have been possible at all without using AI.

ada1981

14 dagen geledendiscuss

Yes I am curious if we will start to see tools that expose this surface more directly or built into the models.

ada1981

15 dagen geledendiscuss

I actually just saw the paper not the one they made. Just googled and found it.

There’s is likely better!

https://www.neuronpedia.org/qwen3.6-27b/jlens?shareId=cmr2kx...

ada1981

15 dagen geledendiscuss

also from our lab is Personality Bench: https://persona.earthpilot.ai/ which runs every frontier model through personality tests.

ada1981

15 dagen geledendiscuss

I run a small AI lab and playground and got super excited about Anthropics paper "Verbalizable Representations Form a Global Workspace in Language Models" (https://transformer-circuits.pub/2026/workspace/index.html)

It talks about how they use a tool they call a Jacobian Lens to view inside the middle layers of LLM while it's working before it commits to a word (token).

I wanted to see if I could get a version of this running on the open models and to my surprise it worked! I ran some experiments with it and build a public facing free tool anyone can use with your own prompts.

Ask the model to describe a symbol of "three curving lines of water" and you can watch "ocean", "sea", and "surf" light up a few layers deeper before it settles on "waves".

You can also edit the internal state. Insert "fire" into the middle layer of the ocean prompt and the answer shifts to something about heat.

For fun / curiosity sake, I also developed way to let the model read its own inner workspace and then decide to suppress or amplify a concept, and run the prompt again.

Interesting finding from running it across models. J-lens beats a plain logit lens on some architectures and does nothing on others, and it isn't about size. A 0.5B Qwen reads better than a 2.8B Pythia. Every Pythia I tried gained basically nothing; the Llama and Qwen models gained a lot. https://lucid.earthpilot.ai/research

This is a 48 hour old project based on emerging research and built on a small model, a small probe set on rented GPUs - but I found it genuinely exciting. The code is open.

I also included a page context "Docent" AI agent you can chat with about whatever you see to help understand what is going on.

Happy to have folks poke around and break it.

I imagine the applications for allowing models to self-reflect / edit internal states can be useful for alignment, confidence, bias detection, etc. and this tool lets you play with the early stages of that.

ada1981

19 dagen geledendiscuss

I discovered an interesting glitch in California DOT law. If you register a vehicle in CA, you can do so with an out of state mailing address (there is no check or verification needed).

Doing so automatically exempts you from CA SMOG because SMOG is a county by county law. If your address is out of state, you have no county smog obligation.

I went to get my vehicle smogged and was told I didn't need to.

ada1981

2 maanden geledendiscuss

I suppose over a long enough time interval where you can absorb the lag the user would receive.

You’d also need monitoring as the networks done reveal exact time stamps on follow / unfollow.

You could build a little toy that automates follow and unfollow and monitors and if two people ran that they could communicate.

Or they could just DM :)

ada1981

2 maanden geledendiscuss

Removing all US financial support for Israel and supporting sanctions and ICJ actions is the only way forward.

ada1981

2 maanden geledendiscuss

Do you mean following and unfollowing in some sort of rapid succession?

ada1981

2 maanden geledendiscuss

[flagged]

ada1981

2 maanden geledendiscuss

Incorrect.