Code Monkey home page Code Monkey logo

llm-arena's Introduction

👋 Hello, I'm IroncladDev

I'm IroncladDev, an eighteen-year-old Fullstack Web Developer who lives out in the country.

I enjoy developing websites, creating content, building with AI, competitive programming, and Brazillian Jiu-Jitsu.

I code because it allows me to express my imagination and creativity. What other programmers develop inspires me to build and inspire others.

llm-arena's People

Contributors

ironcladdev avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar

Forkers

tusharxoxoxo

llm-arena's Issues

Rebrand

Something like LLMArena?

TBD

Address "multiple jotai instances" warning

Getting this error a bunch

Detected multiple Jotai instances. It may cause unexpected behavior with the default store. https://github.com/pmndrs/jotai/discussions/2044

Might migrate away from jotai and just use context providers

Better emails, notify contributors of suspensions/model removals

Currently the emails are dynamic templates in Sendgrid and terribly designed. Would be nice to bring them in here as HTML and send them through the sendgrid API.

If a contributor gets suspended for abuse, if a model gets removed from the platform for abuse, or if one gets approved/rejected, we should send the contributor an email / notification.

Auth wrapper

try-catches look super ugly, repetitive.

Tons of imports for auth 'n stuff.

Add optional 'note' field to Field

Some metrics have labels like "0-shot", "3-shot", etc.
Some HumanEval tests are only in one language or something.

There needs to be a way to fix this.

Screenshot 2024-02-18 at 6 00 00 AM

Permalinks

Allow people to share a link that already has specific LLMs compared side-by-side

Export widget as image

Would be nice if you could export a widget as an image, sort of like how ray.so does it.

A dialog/sheet component would be ideal for this.

Mobile Improvements

Just a couple suggestions for improving the mobile experience.

  1. On the main dashboard, consider giving the main container a dynamic height to prevent unnecessary scrolling (support for dvh was just added in tailwind 3.4)
  2. Update the filters for mobile. Currently, they're a bit crowded. Proposed change: hide the labels (the controls seem clear enough) and move the toggle to a popover on the right hand side.

Screenshot:
WhatsApp Image 2024-02-17 at 17 56 37

Change/edit requests

Add a third type of vote called a "change request" that changes an LLM's vote count but doesn't apply an approval/rejection status.

Change requests get bumped to the top and can be voted on. If the change request gets approved, apply the changes to the LLM (e.g. adding/editing/deleting a metadata field)

Redo number formatting

It's terrible right now, and the input sucks and consists of a ton of components and javascript.

Need to fix, possibly with a third-party library or just make a better implementation.

Add a CI pipeline

If contributors make PRs, there has to be some CI pipelines to ensure everything runs right

Design tables a bit better

With #5, I'll have to space tables out a bit to add additional notes (for benchmarks and stuff)

Boolean tables and string tables look really bad right now.

Search length cap caching

If no results are found with the term "abc", "abcd" should not trigger a search.

Apply in:

  • finding a metadata property
  • finding an LLM (public)
  • finding an LLM (contributors/admins)

Migrate from Prisma to Drizzle

Replit has been having a ton of trouble with Prisma. Found a workaround that you could use NPM to get it to work but not a scalable solution for the future.

OpenSSL versioning and prisma client engine is real wonky on Replit's nix env.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.