Model Overview

The psychopenguin/indian_legal_Phi-3-mini-4k-instruct is a 3.8 billion parameter model from Microsoft's Phi-3 family, specifically the Mini variant with a 4K (4096 token) context length. It has undergone supervised fine-tuning (SFT) and direct preference optimization (DPO) to enhance instruction following and safety. The model demonstrates robust performance in common sense, language understanding, math, code, and logical reasoning benchmarks among models under 13 billion parameters.

Key Capabilities

• Strong Reasoning: Excels in mathematical and logical reasoning tasks.
• Instruction Following: Improved instruction adherence and structured output generation due to post-training updates.
• Efficient Performance: Designed for memory/compute-constrained and latency-bound environments.
• Chat Format Optimized: Best suited for prompts using a chat-based format, supporting multi-turn conversations and system tags.

Intended Uses

This model is suitable for broad commercial and research applications in English, particularly where strong reasoning and efficiency are critical. It serves as a building block for generative AI features and is intended to accelerate research in language models. While it shows strong reasoning, its smaller size means it has limited world knowledge, suggesting augmentation with external tools like search engines for factual recall.