Scaling to Tens of Thousands of Test-Time Iterations with Loop-Native Attention Residuals

Pengxiang Li, Dilxat Muhtar, Di He, Guinan Su, Lu Yin, Shiwei Liu

Adds a learned running summary of previous recurrent states so that each update can retain useful earlier work. Uses content-based weighting and temporal decay with constant aggregation memory, and tests whether puzzle-solving improves over much longer inference runs.

First public: October 8, 2026 · Added to this list: October 18, 2026