This is a grant covering work Celeste did as part of MATS with me (during the period pre MATS where MATS isn't able to cover stipends). I think Celeste did great work here, most notably working on a forthcoming paper I really like training activation oracle variants that verbalize the contents of a LoRA to explain the difference between two models. I have an obvious conflict of interest here, but I think this is clearly above the bar.
1
Interpretability research feb-jun
ProposalGrant
Closes September 15th, 2026$15,000raised
$15,000minimum funding
$15,000funding goal
Project summary
Worked on a bunch of meta model stuff: LoRAcles and improving AOs
What are this project's goals? How will you achieve them?
/
How will this funding be used?
/
Who is on your team? What's your track record on similar projects?
/
What are the most likely causes and outcomes if this project fails?
/
How much money have you raised in the last 12 months, and from where?
some mats funding