arXiv cs.CL
· Papers
Is Convergence Inevitable? Tracing Output Homogeneity Back to Base Models
arXiv:2608.11426v1 Announce Type: new Abstract: The lack of diversity in LM content is widely attributed to the alignment process, but how and where exactly in the pipeline this collapse begins is unknown. We argue that output homogeneity is likely learned during the pretraining phase, and only emph{revealed} or magni