prithivMLmods/Phi-4-Empathetic

Phi-4 Empathetic [ Responsible Reasoning & Emotional Thought Generation ]

[Phi-4 Empathetic finetuned] from Microsoft's Phi-4 is an advanced open model built upon a blend of high-quality synthetic datasets, data from filtered public domain websites, and carefully selected academic resources. It excels at responsible human-like reasoning, empathetic dialogue, and emotional thought generation. The model is designed to engage in nuanced, thoughtful conversations, with outputs that can include special characters and emojis for expressive communication. 🌟

Phi-4 Empathetic employs a sophisticated safety post-training approach, leveraging both open-source and proprietary datasets. Safety alignment is achieved using a combination of SFT (Supervised Fine-Tuning) and DPO (Direct Preference Optimization), targeting responsible interaction and emotional awareness in diverse contexts.

---

Dataset Info

Phi-4 Empathetic is fine-tuned on a carefully curated dataset tailored for empathetic and responsible reasoning tasks. The dataset incorporates the Chain of Thought (CoT) methodology, emphasizing logical reasoning, emotional nuance, and step-by-step thought processes. Additionally, it includes data optimized for generating responses that resonate with human emotions, making it ideal for:

• Emotional Support Applications 🤗
• Responsible Conversations 💬
• Thoughtful Problem-Solving 🧠

---

Run with Transformers

# pip install accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

tokenizer = AutoTokenizer.from_pretrained("prithivMLmods/Phi-4-Empathetic")
model = AutoModelForCausalLM.from_pretrained(
    "prithivMLmods/Phi-4-Empathetic",
    device_map="auto",
    torch_dtype=torch.bfloat16,
)

input_text = "Can you share some words of encouragement for someone feeling down?"
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")

outputs = model.generate(**input_ids, max_new_tokens=32)
print(tokenizer.decode(outputs[0]))

You can ensure correct formatting for empathetic dialogue by using tokenizer.apply_chat_template as follows:

messages = [
    {"role": "user", "content": "Can you share some words of encouragement for someone feeling down?"},
]
input_ids = tokenizer.apply_chat_template(messages, return_tensors="pt", return_dict=True).to("cuda")

outputs = model.generate(**input_ids, max_new_tokens=256)
print(tokenizer.decode(outputs[0]))

---

Intended Use

The Phi-4 Empathetic model is optimized for applications that require thoughtful and emotionally aware interactions. Below are some suggested use cases:

1. Emotional Support & Counseling 💖

   • Providing thoughtful responses to users seeking emotional encouragement or advice.
   • Generating empathetic messages for mental health and well-being applications.

2. Responsible Dialogue Generation 🗣️

   • Engaging in nuanced conversations with a focus on fairness, safety, and ethical considerations.
   • Ensuring that interactions remain respectful and aligned with safety guidelines.

3. Creative Writing Assistance ✍️

   • Helping users craft emotionally engaging content, including stories, poems, and personal messages.
   • Assisting in generating content enriched with special characters and emojis for expressive communication.

4. Educational Tools 🎓

   • Offering step-by-step explanations with an empathetic tone for better understanding.
   • Generating thoughtful Q&A responses for various subjects.

5. Customer Support 🤝

   • Automating empathetic responses to customer queries.
   • Handling emotionally sensitive customer service interactions with care.

6. Social Media Engagement 📱

   • Generating creative, engaging, and emotionally resonant posts for social media platforms.
   • Providing personalized message suggestions enriched with emojis and special characters.

---

Limitations

While Phi-4 Empathetic is highly capable, it has certain limitations users should be aware of:

1. Bias and Fairness:
   Despite extensive safety alignment, biases may still emerge in the model’s responses. Users should exercise discretion, particularly in sensitive contexts.

2. Emotional Nuance:
   The model may occasionally misinterpret the emotional tone of a prompt, leading to less relevant or inappropriate responses.

3. Real-Time Knowledge:
   The model's knowledge is based on the data it was trained on and does not include real-time or post-training updates. It may not reflect recent events or changes in knowledge.

4. Safety and Harmlessness:
   Although the model is aligned with safety standards, there may still be cases where outputs require human oversight to ensure appropriateness.

5. Resource Requirements:
   Running the model efficiently may require significant computational resources, especially in large-scale or real-time applications.

6. Ethical Considerations:
   The model must be used responsibly, avoiding any malicious applications such as generating harmful content or spreading misinformation.

7. Domain-Specific Limitations:
   While it performs well in general-purpose tasks, it may need further fine-tuning for highly specialized domains, such as legal, medical, or financial applications.

---

Special Features

1. Emojis & Special Characters 🎉💡
   The model can generate responses with emojis and special characters for expressive communication, making it ideal for social media and personal messaging applications.

2. Human-Like Reasoning 🧠
   Fine-tuned for responsible reasoning and empathetic dialogue, it excels at generating thoughtful and human-like responses.

3. Advanced Safety Alignment 🔒
   The model employs iterative SFT and DPO techniques to ensure that its outputs are helpful, harmless, and aligned with ethical standards.