BWhat changed
Scored corpus tokens with a reference model and focused training loss on higher excess-loss tokens, evaluating token-selection effects in math and general language-model training.
Scientific scope
Scored corpus tokens with a reference model and focused training loss on higher excess-loss tokens, evaluating token-selection effects in math and general language-model training. Token/data selection has predecessors. Reference-model training/scoring costs remain; fewer selected training tokens are not automatically equivalent end-to-end compute savings. Unrelated to Reka's2026Rho-1 multimodal system; no grading/ranking scores assigned by WRS.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Token/data selection has predecessors. Reference-model training/scoring costs remain; fewer selected training tokens are not automatically equivalent end-to-end compute savings. Unrelated to Reka's2026Rho-1 multimodal system; no grading/ranking scores assigned by WRS.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction