From 9be38693f46ff766a4fb93378c98949acf910249 Mon Sep 17 00:00:00 2001
From: PhilipMay <philip@may.la>
Date: Fri, 29 Dec 2023 15:05:25 +0100
Subject: [PATCH] add llm

---
 source/machine-learning/llm.md | 7 +++++++
 1 file changed, 7 insertions(+)
 create mode 100644 source/machine-learning/llm.md

diff --git a/source/machine-learning/llm.md b/source/machine-learning/llm.md
new file mode 100644
index 0000000..9d05473
--- /dev/null
+++ b/source/machine-learning/llm.md
@@ -0,0 +1,7 @@
+# LLM
+
+## Base Knowledge
+- Direct Preference Optimization (DPO)
+  - Paper: <https://arxiv.org/abs/2305.18290>
+  - <https://plainenglish.io/community/direct-preference-optimization-dpo-a-simplified-approach-to-fine-tuning-large-language-models>
+  - <https://huggingface.co/blog/dpo-trl>