2026-08-09T06:30:20Z https://dcpapers.dublincore.org/oai

oai:dcpapers.dublincore.org:952493684 2024-12-20 dcmi-2024

Synthetic Signal Identification in LLM Datasets Hahn, Jim This research addresses the quality of training data in LLMs using methods from signaling theory and the talk page metadata of Wikipedia articles. The significance of the method is to lower the cost of information quality assessment in datasets. Natural language processing on metadata text generated sentiment, reading complexity, and self-reference scores as contributions to the computationally derived signals. Results showed that it is possible to understand indicators of information quality using textual computation over the metadata in article pages. Dublin Core Metadata Initiative 2024-12-20 Text application/pdf text/html https://doi.org/10.23106/dcmi.952493684 https://dcpapers.dublincore.org/article/952493684 Dublin Core Metadata Initiative Conference Proceedings en https://creativecommons.org/licenses/by/4.0/