You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
$6,000 left to reach the funding goal.
We are AI Interpretability @ Illinois (https://interpretability.web.illinois.edu/), a cross-lab research group based at UIUC with 5 faculty advisors and 30+ student members working on technical AI safety, especially through making the black-box LLM interpretable, auditable and trustworthy. In our first year, we produced 7 publications on several subtopics (https://interpretability.web.illinois.edu/research-projects/), and 5 are under review to a conference at this moment (4 to ARR, 1 to AISTATS), and 8 concurrent projects are under their way and will be submitted to other top AI conferences such as ICML, Neurips, ICLR, etc. We want to ask for your help to financially support us student authors to travel to the AI conferences within the US when a paper gets accepted (usually only the first author per paper), especially at this difficult moment when school fundings for graduate students are very limited to support the conference travel.
To support student researchers cover the expense of a conference travel; with the funding raised from Manifund, I can reimburse each author's travel cost based on the actual expense.
I'd expect to see more student authors go to the conference in person with the funding who typically didn't attend even a paper was accepted.
Support student researchers cover the expense of a conference travel, via reimbursement of their actual travel cost within the US with an limit (e.g. $500 maximum per person).
We have 5 faculty members and 30+ student researchers at this moment: https://interpretability.web.illinois.edu/people/
The track records include the follows:
(a) Publications: In our first year, we produced 7 publications on several subtopics (https://interpretability.web.illinois.edu/research-projects/), and 5 are under review to a conference at this moment (4 to ARR, 1 to AISTATS), and 8 concurrent projects are under their way and will be submitted to other top AI conferences such as ICML, Neurips, ICLR, etc.
(b) Resources we secured: We secured 45k+ GPU-hours on NCSA Delta and $20k+ in credits via YC University Project with me being the PI, and those resources are well used in our projects with proper acknowledgement at each paper.
(c) Group running: Current members typically meet bi-weekly within the project they are working on. We also offer a seminar at UIUC: CS591 Biologically Plausible AI (https://publish.illinois.edu/cs591-bai/).
(d) Visibility: I got invited to present work from AI Interpretability at Citadel’s GQS PhD Colloquium 2026.
At this moment, it's very unlikely that this project would fail, as we have established solid research management cycles, with active researcher members involved and abundant computing resources needed for our experiments. This pipeline works well and has output 7 publications.
I haven't raised any money before. I only secured computing credits.