kaal:claim:7598518-027

Claim. Reputation is settled once by key, after the generation-0 fit on real data and before generation 1, and no candidate that enters training ever updates a reputation.

Type. design · Support. stated

Holds when

• The design uses only keyed settlement, although the parent rule also covers unkeyed events.
• This removes reputation dynamics during the recursion, so the weights are fixed throughout the run.

Source quote

This study uses only keyed settlement, once, after the generation-0 fit on real data and before generation 1. No candidate that enters training ever updates a reputation.

From. Wulf A. Kaal, Can Reputation Prevent Model Collapse? A Preregistered Test of Reputation-Weighted Data Filtering (2026). Full text.

Cite as. Wulf A. Kaal, Can Reputation Prevent Model Collapse? A Preregistered Test of Reputation-Weighted Data Filtering (2026). SSRN: https://ssrn.com/abstract=7598518

Verify. sha256 of source PDF cd22d148b2b55c2e27f087dcaa61cdf44ad7022b400b355c31a9cd7e43a31749 at https://wulfkaal.com/papers/7598518.pdf

Content hash. 7e50d4839e4f6fe08904c8d65555a6a6292372dff0579d4898b1f5789c11f1e8