kaal:claim:7598518-029

Claim. Three validator families share a mean accuracy of 0.75 while differing in structure: one is homogeneous at 0.75, one has each validator reliable at 0.90 on two hash-drawn tags and at 0.60 on the other two so that domain-scoped reputation has something to find, and one has five validators at 0.90 and fifteen at 0.70, a competent minority.

Type. design · Support. stated

Holds when

• The validators are simulated and err independently given their accuracy, which depends only on the tag.
• The families are a property of the design, fixed before any run.

Source quote

Family U is homogeneous at 0.75. In family H each validator is reliable at 0.90 on two hash-drawn tags and at 0.60 on the other two, so domain-scoped reputation has something to find. Family Mn has five validators at 0.90 and fifteen at 0.70, a competent minority. The three share a mean accuracy of 0.75.

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. d8a3ba1654989f1cf7dc676997fb7fa29ea53e035585f92db53bb2ac8748616d