appetrosyan.github.io/posts/benchmark-llm

Preview meta tags from the appetrosyan.github.io website.

Linked Hostnames

4

Search Engine Appearance

Google

https://appetrosyan.github.io/posts/benchmark-llm

Benchmarking local LLMs

I've dealt with an ego-maniac working on LLMs when I started my industry career in 2020. Yes, I'm relatively new in the programming scene, at least with regards to having commercial experience. This left a sour taste in my mouth, leading to me being very cautious of using LLMs in general, and using the mainstream LLMs, such as ChatGPT, Claude and to some extent Deepseek with extreme caution. I do however, find that if the process of using these results in an improvement in one's life, they are a tool that needs exploring, and improving. Realistically, I found a set up that is good, at least in my book: Emacs with ellama and a local collection of LLMs.



Bing

Benchmarking local LLMs

https://appetrosyan.github.io/posts/benchmark-llm

I've dealt with an ego-maniac working on LLMs when I started my industry career in 2020. Yes, I'm relatively new in the programming scene, at least with regards to having commercial experience. This left a sour taste in my mouth, leading to me being very cautious of using LLMs in general, and using the mainstream LLMs, such as ChatGPT, Claude and to some extent Deepseek with extreme caution. I do however, find that if the process of using these results in an improvement in one's life, they are a tool that needs exploring, and improving. Realistically, I found a set up that is good, at least in my book: Emacs with ellama and a local collection of LLMs.



DuckDuckGo

https://appetrosyan.github.io/posts/benchmark-llm

Benchmarking local LLMs

I've dealt with an ego-maniac working on LLMs when I started my industry career in 2020. Yes, I'm relatively new in the programming scene, at least with regards to having commercial experience. This left a sour taste in my mouth, leading to me being very cautious of using LLMs in general, and using the mainstream LLMs, such as ChatGPT, Claude and to some extent Deepseek with extreme caution. I do however, find that if the process of using these results in an improvement in one's life, they are a tool that needs exploring, and improving. Realistically, I found a set up that is good, at least in my book: Emacs with ellama and a local collection of LLMs.

  • General Meta Tags

    13
    • title
      Benchmarking local LLMs · Aleksandr Petrosyan
    • charset
      utf-8
    • viewport
      width=device-width,initial-scale=1
    • color-scheme
      light dark
    • author
      Aleksandr Petrosyan

  • Open Graph Meta Tags

    6
    • og:url
      https://appetrosyan.github.io/posts/benchmark-llm/
    • og:site_name
      Aleksandr Petrosyan
    • og:title
      Benchmarking local LLMs
    • og:description
      I've dealt with an ego-maniac working on LLMs when I started my industry career in 2020. Yes, I'm relatively new in the programming scene, at least with regards to having commercial experience. This left a sour taste in my mouth, leading to me being very cautious of using LLMs in general, and using the mainstream LLMs, such as ChatGPT, Claude and to some extent Deepseek with extreme caution. I do however, find that if the process of using these results in an improvement in one's life, they are a tool that needs exploring, and improving. Realistically, I found a set up that is good, at least in my book: Emacs with ellama and a local collection of LLMs.
    • og:locale
      en

  • Twitter Meta Tags

    3
    • twitter:card
      summary
    • twitter:title
      Benchmarking local LLMs
    • twitter:description
      I've dealt with an ego-maniac working on LLMs when I started my industry career in 2020. Yes, I'm relatively new in the programming scene, at least with regards to having commercial experience. This left a sour taste in my mouth, leading to me being very cautious of using LLMs in general, and using the mainstream LLMs, such as ChatGPT, Claude and to some extent Deepseek with extreme caution. I do however, find that if the process of using these results in an improvement in one's life, they are a tool that needs exploring, and improving. Realistically, I found a set up that is good, at least in my book: Emacs with ellama and a local collection of LLMs.

  • Link Tags

    13
    • apple-touch-icon
      /images/apple-touch-icon.png
    • apple-touch-icon
      /images/apple-touch-icon.png
    • canonical
      https://appetrosyan.github.io/posts/benchmark-llm/
    • icon
      /img/favicon.svg
    • icon
      /img/favicon-32x32.png

Links

11