park-2025-gemma-2b-specifications

IN premise — summaries/2026/08/24/park-2024-categorical-hierarchical-concepts-sR-references-chunk-1.md

Created 2026-08-25T02:58:25+00:00

Gemma-2B has 2 billion parameters, was pretrained on 3 trillion tokens, has a 256K vocabulary, and uses a 2048-dimensional representation space.

Summary

This records the core size and training profile of the Gemma-2B model so that any downstream comparison, capacity check, or resource estimate in the system has a verified factual starting point. It matters because claims about what the model can do, how much memory it needs, or how it compares to other models all depend on these numbers being correct.