ripple-edits-models-evaluated
IN premise — summaries/2026/08/24/cohen-2023-ripple-effects-s5-experiments.md
Created 2026-08-25T02:57:57+00:00
RIPPLE EDITS evaluates four open-source LMs (GPT-2 XL 1.5B, GPT-J 6B, LLaMA 7B, GPT-NeoX 20B) and GPT-3 (text-davinci-003, 175B) across ROME, MEMIT, MEND, and ICE methods.
Summary
The RIPPLE EDITS study tested its editing approach against five language models ranging from 1.5 to 175 billion parameters, comparing it alongside four existing editing methods (ROME, MEMIT, MEND, and ICE). This matters because the evidence base is broad enough to support conclusions that generalize across model sizes and competing techniques, rather than reflecting a single favorable setup.