Controlled framework for iterative model-guided candidate generation, validation, evaluation, and selection with reproducible accounting.
-
Updated
Sep 16, 2026 - Python
Controlled framework for iterative model-guided candidate generation, validation, evaluation, and selection with reproducible accounting.
Looped-transformer experiments with bounded RSI-style iterative adaptation, recurrent depth, and post-selection held-out evaluation.
To associate your repository with the iterative-adaptation topic, visit your repo's landing page and select "manage topics."