Inside the Loop: A Mechanistic Study of Weight-Tied Transformers on Depth-Bound Algorithmic Tasks
Trains looped and untied transformers to follow paths through a shuffled permutation graph without writing intermediate hops. Uses probes and activation patching to test when intermediate nodes become available and whether later computation uses them.
First public: · Added to this list: October 22, 2026
Explore this approach
Recurrent reasoning — Repeat parts of the model or keep intermediate information in memory while working through a problem. Some methods also learn how many passes to run.
Analysis and evaluation — Measure how well these methods work, investigate what their internal computation does, and identify where they fail. This includes experiments, studies of model internals, and mathematical results.