AWhat changed
Used internal states from a pretrained bidirectional language model as token-context representations, learning task-specific layer mixtures to augment downstream NLP models.
Scientific scope
Used internal states from a pretrained bidirectional language model as token-context representations, learning task-specific layer mixtures to augment downstream NLP models. Inherits LSTMs and earlier contextual language/translation representations. Distinctive deep layer mixing is not first invention of context-dependent embeddings or modern masked Transformer pretraining. Original experiments freeze pretrained LM weights and train task modules; source benchmark gains are not universal transfer guarantees.
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 LSTMs and earlier contextual language/translation representations. Distinctive deep layer mixing is not first invention of context-dependent embeddings or modern masked Transformer pretraining. Original experiments freeze pretrained LM weights and train task modules; source benchmark gains are not universal transfer guarantees.
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