Abhay Koul

Abhaykoul

AI & ML interests

None yet

Recent Activity

upvoted a collection 3 days ago
HelpingAI2.5
updated a collection 4 days ago
HelpingAI2.5
liked a model 4 days ago
OEvortex/HelpingAI2.5-10B

Articles

Organizations

Posts 3

view post
Post
2730
Introducing HelpingAI2-9B, an emotionally intelligent LLM.
Model Link : OEvortex/HelpingAI2-9B
Demo Link: Abhaykoul/HelpingAI2

This model is part of the innovative HelpingAI series and it stands out for its ability to engage users with emotional understanding.

Key Features:
-----------------

* It gets 95.89 score on EQ Bench greather than all top notch LLMs, reflecting advanced emotional recognition.
* It gives responses in empathetic and supportive manner.

Must try our demo: Abhaykoul/HelpingAI2
view post
Post
2559
# HelpingAI 9B: Cutting Edge Emotionally Intelligent AI

If you have ever felt that AI not understand your emotions or you not get human like fell while taking to him than this blog is for you!
In this BlogPost we will be exploring [HelpingAI 9B](https://huggingface.co/spaces/Abhaykoul/HelpingAI-9B) is an Highly Emotionally Intelligent AI which beated all top notch ai like GPT4o, GPT4, Claude3 Opus on EQ-Bench.

## What is HelpingAI 9B?
![image/png](https://cdn-uploads.huggingface.co/production/uploads/6612aedf09f16e7347dfa7e1/FrNhr3WhMhvvD-dNplHxZ.png)
HelpingAI-9B is the fine-tuned Llama2 model crafted for emotionally intelligent conversations. This model excels in empathetic engagement, offering understanding and support through dialogue spanning various topics and situations. Its goal is to serve as a supportive AI companion, adept at resonating with users' emotions and communication requirements.

## Method
We gathered a large volume of high-quality human chat data, which was then filtered and refined to create three types of datasets:
1. DPO - Initially, we trained the AI on a substantial DPO dataset to grasp human conversation patterns, enabling it to discern which types of output to generate and which to avoid.
2. Alpaca - Subsequently, we trained it on an Alpaca-type dataset to enhance its human-like responses.
3. SFT - Finally, once the AI fully comprehended human interactions, we trained it on an SFT dataset to broaden its knowledge base.

## Evaluation

![By/KingNish/Banchmark/png](https://cdn-uploads.huggingface.co/production/uploads/6612aedf09f16e7347dfa7e1/xvS57q5kU9f3AX-O-al0k.png)

## Conclusion:
HelpingAI is a large step toward understanding human emoions and give response like them. It can help us lot in making AI better for NLP.
Thanks!

Model Link: - OEvortex/HelpingAI-9B

Demo link: - https://huggingface.co/spaces/Abhaykoul/HelpingAI-9B