Lossfunk Residency Project Proposal
Be specific. Not "I want to explore mechanistic interpretability" but "I want to find out whether attention heads in 2-layer transformers learn the same circuits across random seeds when trained on modular arithmetic." If you can't write it at this level of precision, keep thinking.
Limit: 100 words
What's the origin? What made you pick this specific project rather than 50 other things you could have done? We want the real story (a paper that bothered you, an observation that wouldn't leave you alone, a hunch you've been sitting on). Generic "I'm interested in AI" answers tell us nothing.
Limit: 100 words
Concrete output, not vague aspiration. Not "a deeper understanding" but a specific thing you'll have: a working prototype, a writeup with specific findings, a benchmark with numbers, a tool other people can use. If you can't describe success concretely, you don't yet know what you're building.
Limit: 100 words
Share one thing you're most proud of — a side project, blog post, essay, repo, tool, thread, anything. Something you built or wrote because you wanted to, not because someone assigned it. Include a link. Then tell us in one sentence why it matters to you.
Link + one sentence