ComfyAI·Sep 10Prior-Data Fitted Networks: The Paper That Became TabPFNThis ICLR paper introduced Prior-Data Fitted Networks (PFNs), the work that later grew into TabPFN. Here’s my reading of it.A response icon1
ComfyAI·Sep 6TabPFN in Under 500 LinesThis one comes from the original TabPFN team, though it isn’t really a paper. It’s a small GitHub project: TabPFN v2 rewritten from scratch…
ComfyAI·Sep 5Stop Retrieving ICL Examples by Input Similarity — Retrieve by the Causal PartIn-context learning works by dropping a few examples into the prompt so the model has something to imitate when it answers. It’s cheap, it…
ComfyAI·Sep 4Can In-Context Learning Actually Learn Tasks It Never Saw in Training?In-context learning is the ability that made large language models feel magical: you show the model a few examples, and it seems to pick up…
ComfyAI·Aug 24How a Few Examples Become One Task VectorWhen you hand a language model a handful of demonstrations and it suddenly “gets” what you want, something has to happen inside the network…
