# FineTune LLM on a behavior strategy

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffinetune-llm-on-a-behavior-strategy%2F155331%3E%5C
**Category:** Beginners\
**Created:** [May 15, 2025, 8:22am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffinetune-llm-on-a-behavior-strategy%2F155331 "2025-05-15T08:22:55Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![ammeni](https://avatars.discourse-cdn.com/v4/letter/a/f17d59/32.png) [@ammeni](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fammeni)%5C
**Post date:** [May 15, 2025, 8:22am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffinetune-llm-on-a-behavior-strategy%2F155331%2F1 "2025-05-15T08:22:55Z")

</div>

Hello,  
I am finetuning an LLM to adopt a specific teaching methodology, I have some papers and articles and a huge book about this methodology. How should i proceed ?  
Thank you

---

<div class="post-metadata">

**Author:** ![John6666](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/john6666/32/27664_2.png) [@John6666](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2FJohn6666)%5C
**Post date:** [May 15, 2025, 2:15pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffinetune-llm-on-a-behavior-strategy%2F155331%2F2 "2025-05-15T14:15:34Z")

</div>

There is a lot to do, but first, you maybe need to decide whether to use a RAG-based approach and then select a base model.  
If you use a RAG-based approach, even smaller models such as 3B to 7B can achieve decent performance.  
If you want to pack knowledge into an LLM alone, you will need a larger model.

> **[The Large Language Model Course](https://huggingface.co/blog/mlabonne/llm-course)**
>
> A Blog post by Maxime Labonne on Hugging Face

> **[The Big Benchmarks Collection - a open-llm-leaderboard Collection](https://huggingface.co/collections/open-llm-leaderboard/the-big-benchmarks-collection-64faca6335a7fc7d4ffe974a)**
>
> Gathering benchmark spaces on the hub (beyond the Open LLM Leaderboard)

---

<div class="post-metadata">

**Author:** ![ammeni](https://avatars.discourse-cdn.com/v4/letter/a/f17d59/32.png) [@ammeni](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fammeni)%5C
**Post date:** [May 15, 2025, 2:34pm UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Ffinetune-llm-on-a-behavior-strategy%2F155331%2F3 "2025-05-15T14:34:23Z")

</div>

Thank you for your response. I think for my case finetuning would be the better option since I already want it to adopt this teaching methodology as a whole personality and my data won’t change and be external, it’s just once and then I will use this finetuned version for other tasks in the future.