What's in a prompt? Language models encode literary style in prompt embeddings
This work provides insights into the sophistication of information processing in language models, with potential applications in authorship attribution and literary analysis, though it is incremental in nature.
The study investigated how large language models encode intangible aspects like literary style in prompt embeddings, finding that short excerpts from different novels separate in latent space independently of next-token predictions and that embeddings from the same author are more entangled than across authors.
Large language models use high-dimensional latent spaces to encode and process textual information. Much work has investigated how the conceptual content of words translates into geometrical relationships between their vector representations. Fewer studies analyze how the cumulative information of an entire prompt becomes condensed into individual embeddings under the action of transformer layers. We use literary pieces to show that information about intangible, rather than factual, aspects of the prompt are contained in deep representations. We observe that short excerpts (10 - 100 tokens) from different novels separate in the latent space independently from what next-token prediction they converge towards. Ensembles from books from the same authors are much more entangled than across authors, suggesting that embeddings encode stylistic features. This geometry of style may have applications for authorship attribution and literary analysis, but most importantly reveals the sophistication of information processing and compression accomplished by language models.