
FL
Fengming Liu
· 1 min read
ResearcharXiv cs.CL
Embedding Cosine Measures Text Length Before It Measures Homogenization
arXiv:2611.21519v1 Announce Type: new
Abstract: Mean pairwise cosine similarity between transformer embeddings is a standard measure of whether online discourse is converging under generative AI. When the embeddings are mean-pooled outputs of an encoder never trained for sentence similarity, this measure mostly tracks text length. In 6.06 million Reddit comments, an uncentered DeBERTa-v3 measure had suggested that only conservative discourse broke toward homogeneity after late 2022. Truncating the same comments from 10 to 150 words raises their mean cosine by eight times the reported break. A placebo that swaps every comment for a random one of equal length reproduces 81% of the asymmetry, and a fixed-content simulation reproduces 78%. The leading principal direction carries 99.9% of raw DeBERTa homogeneity, and once length is handled the asymmetry disappears under every encoder. On 575,000 Stack Exchange answers the measure reports a post-ChatGPT homogenization of technical sites that the placebo fully reproduces, although answers grew longer. A census of 31 studies locates the risk where anisotropic encoders meet time-series or length-unequal designs. A reverse control on a well-known experiment shows that centering with renormalization can reverse a genuine effect. I propose a low-cost diagnostic protocol. Corrected measures show that the two political camps converged between 2019 and 2022 and stopped converging before ChatGPT. Cross-camp replies engage their targets more closely than same-camp replies.
Original source
This story was published by arXiv cs.CL and written by Fengming Liu. SyncAI.news shows a preview; the complete article is on the publisher's site.
Read the full story on arxiv.org


