Recursive Language Models and Semantic Pattern Matching
Experiments with recursive inference over deeply nested, multi-schema data sources.
At the end of 2025, Alex L. Zhang, Tim Kraska, and Omar Khattab published a paper, Recursive Language Models.
Since then, I've been experimenting with it for the use case of semantic pattern matching in deeply nested, multi-schema data sources.
I was excited to find Recursive Agent Optimization yesterday, which demonstrates this same pattern on models trained to work with the RLM scaffold.