From d08b0e4a2eedd2bb89077845abef9a4c37f26de0 Mon Sep 17 00:00:00 2001
From: "wassname (Michael J Clark)" <1103714+wassname@users.noreply.github.com>
Date: Mon, 9 Mar 2026 06:43:34 +0800
Subject: [PATCH] Update README.md

---
 README.md | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/README.md b/README.md
index b0f0238..04462b8 100644
--- a/README.md
+++ b/README.md
@@ -91,3 +91,6 @@ If this research contributes to your work, please acknowledge it by citing:
   commit = {}
 }
 ```
+
+UPDATE:
+I think a better approach might be to use influence functions such as: https://joemelko.github.io/blog.html?post=TopicReweighting these essentially meta learn a function to see how influential data would be if you did a larger and more expensive train.