model-capacity-ranking-shen-2023
IN premise — summaries/2026/08/24/shen-2023-icl-not-gd-s8-icl-demonstrations.md
Created 2026-08-25T02:58:33+00:00
The model capacity ranking used in Shen et al. (2023) experiments is LLaMA (7B) > GPT-J (6B) > GPT-Neo (2.7B) > GPT2-XL (1.5B).
Summary
This establishes the specific model-size ordering the Shen et al. 2023 experiments relied on, running from a 7B-parameter LLaMA at the top down to a 1.5B GPT2-XL at the bottom. It matters because every claim in that study about how model capacity affects performance is anchored to this particular ranking, and the system needs it locked in to interpret those results consistently.