BWhat changed
Unified heterogeneous moderation taxonomies into instruction/query/document yes-no tasks, training on curated and contrastively generated text/image data and using yes/no logits for policy-conditioned classification.
Scientific scope
Unified heterogeneous moderation taxonomies into instruction/query/document yes-no tasks, training on curated and contrastively generated text/image data and using yes/no logits for policy-conditioned classification. Inherits pretrained Ministral and supervised classification/merging. A continuous classifier output is not calibrated universal risk, guaranteed safety, compliance or immunity to adversarial inputs. Named core contributors are source credit, not independently audited individual roles.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits pretrained Ministral and supervised classification/merging. A continuous classifier output is not calibrated universal risk, guaranteed safety, compliance or immunity to adversarial inputs. Named core contributors are source credit, not independently audited individual roles.
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