InfiLoop Preprint: Loop-Native Residuals Boost 7M Model to 97.9% Accuracy on Sudoku

The InfiLoop method enables a 7M-parameter model to reach 97.9% accuracy on Sudoku-Extreme and continue improving with over 20,000 test-time iterations.

Background: Existing looped Transformers see reduced reasoning accuracy as iterations increase, because noisy state updates overwrite correct intermediate deductions and even undo completed solutions, making error correction difficult for later loops.

Conclusion: The preprint introduces InfiLoop, a loop-native residual connection combining content-based weighting with learned temporal decay. It learns which past computations to retain and how much to accept from new updates, suppressing unreliable proposals while preserving useful states.

Boundary: Results are self-reported by the authors and not yet independently reproduced; validation is limited to specific reasoning tasks like Sudoku and ARC-AGI-2, with no stated performance on general language models.