Luna Nova

@lunnova.dev

she/her | posts at https://lunnova.dev | code at https://github.com/LunNova Give some of your money to effective charities if you're well off!

That'll do it. aint go agent getting into my infra any time soon! Would like to see GPT-6 try to get past this...

Bild

Steering vectors can work as activation oracles! Inspired by @isolyth.dev 's recent work on instruct vectors, I've ported the same approach to create Activation Oracles. Seems to work surprisingly well given the param deficit, nearly matching Activation Oracle LoRA performance.

Figure showing a vector based and LoRA based activation oracle's results at extracting a taboo word. ≈43% success for the vector, and ≈50% success for the LoRA.

Anyone have a favorite model of burnout, executive function, akrasia, exhaustion, inertia, spoons, mana etc. that treats activation energy as a minimal energy level without treating it as spent, or otherwise doesn't use a global energy meter? n=1 anecdotes fine!

my shader export compression is mightily broken and the effect is eerily reminiscent of trying to make out details in a night sky obscured by light pollution, except the sky now has intermittent complex transforms darting across its surface

The OLMo fully open foundation models are probably not going to get new revisions. Pretty sad as they were one of the few true open source options - mostly 'open weights' models don't share training recipes or data and are more like freeware.

GeekWire excerpt: "In addition, while all Ai2 programs for 2026 are fully funded, these people said, FFST is moving from providing Ai2 with overall annual funding to a proposal-based process, with future support expected to favor real-world applications of AI over building open-source foundation models. The shift helps explain the departures of researchers focused on model development."

The words 'over building open-source foundation' are highlighted in green.