Skip to content

RL ChatMate is an open-source project aimed at improving chatbot interactions using state-of-the-art reinforcement learning techniques. By fine-tuning language models with human feedback, we enhance response quality, coherence, and engagement.

License

Notifications You must be signed in to change notification settings

Q-Aware-Labs/RL-ChatMate

Repository files navigation

RL-ChatMate

******* 🚧⚠️ PROJECT IS UNDER CONSTRUCTION ⚠️🚧 *********

RL ChatMate is an open-source project aimed at improving chatbot interactions using state-of-the-art reinforcement learning techniques. By fine-tuning language models with human feedback, the main goal is to enhance response quality, coherence, and engagement.

The project is organized into the following sections:

README.md: Read about the project and its main goal. About tokenization: a brief description of the tokenization process. Dataset tokenization: Chi-Square Test: The execution of the chi-square test and hypothesis validation.

The dataset

For this project I will be using the CommonsenseQA dataset.

CommonsenseQA is a new multiple-choice question answering dataset that requires different types of commonsense knowledge to predict the correct answers . It contains 12,102 questions with one correct answer and four distractor answers. The dataset is provided in two major training/validation/testing set splits: "Random split" which is the main evaluation split, and "Question token split", see paper for details.

Originals autors: Alon Talmor, Jonathan Herzig, Nicholas Lourie, Jonathan Berant

From: https://arxiv.org/abs/1811.00937 arXiv:1811.00937v2 [cs.CL]

Official website:

https://www.tau-nlp.sites.tau.ac.il/commonsenseqa

About

RL ChatMate is an open-source project aimed at improving chatbot interactions using state-of-the-art reinforcement learning techniques. By fine-tuning language models with human feedback, we enhance response quality, coherence, and engagement.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published