# kaal:claim:7456999-044

**Claim.** Format D would train on arguments accompanying a staked vote change, but that use is untested.

**Type.** failure  **Support.** asserted

**Holds when.**

- the proposed Debate-Informed Revision training format

**Source quote.**

> Format D would train a model on the arguments that accompanied a staked vote change. That use is untested.

**From.** Wulf A. Kaal, *Reinforcement Learning from Reputation Feedback* (2026), III.D.2 The Vote-Change Delta as a Novel Training Signal, page 15

**Cite as.** Wulf A. Kaal, Reinforcement Learning from Reputation Feedback (2026). SSRN: https://ssrn.com/abstract=7456999

**Verify.** sha256 of source PDF `01016fbbcc25cad205cebd59ea686d8dfee8ae9b632095d86ba6c9a900456a4b` at https://raw.githubusercontent.com/wulfkaal/Academic-Papers/main/papers/pdf/Kaal%20-%202026%20-%20Reinforcement%20Learning%20from%20Reputation%20Feedback.pdf

**Topics.** ai-and-agents, research-methods, reputation

**Canonical form.** This markdown file is the canonical hashed representation of the claim. Its sha256 is the content hash used for attestation.