LLM Classification Finetuning

Overview

This competition challenges you to predict which responses users will prefer when interacting with different large language models (LLMs). You'll use a dataset of conversations from Chatbot Arena to build a machine learning model. Your work will help improve LLMs to better align with human preferences, making chatbots more user-friendly.

Requirements

This is a 'Getting Started' competition, which means it's great for beginners and has a rolling timeline with no cash prizes. You can participate individually or in teams. Submissions must be made through Kaggle Notebooks that run within 9 hours on CPU or GPU and have internet access disabled. You can use freely available external data and pre-trained models. Your submission file must be a CSV named 'submission.csv' with 'id' and probabilities for 'winnermodela', 'winnermodelb', and 'winner_tie'.

Prizes

This competition is a 'Getting Started' competition and does not offer cash prizes. Instead, it focuses on learning and gaining experience with Kaggle's platform and machine learning concepts.