llaa33219/BokantLM0.1-135M-SmolLM2-135M-LoRA-claude-3.7-sonnet

BokantLM0.1-135M-claude-3.7-sonnet

BokantLM – "Small but Supreme in Its Domain"

BokantLM is not a general-purpose model that tries to do everything well.
Instead, it is an ultra-lightweight LLM designed to focus on a single domain, delivering the highest possible efficiency and performance in that area.

---

Overview

• Model Name: BokantLM0.1-135M-claude-3.7-sonnet
• Base Model: SmolLM2-135M
• Fine-tuning Dataset: reedmayhew/claude-3.7-sonnet-reasoning

---

Philosophy

While most LLMs aim for versatility by learning across many fields,
BokantLM is built to achieve top efficiency and performance within a specific domain.

This model has been fine-tuned to mimic Claude 3.7 Sonnet's tone, reasoning style, and linguistic habits, and despite its small parameter count, it produces responses that are reasonably similar to those of Claude 3.7 Sonnet.

The reason I created this model

I became curious about what would happen if a very small-parameter model tried to mimic Claude 3.7 Sonnet (Thinking) —specifically, how much reasoning capability it could achieve and how much its performance might improve—so I built this model.

---

Future Plans

• ✅ Imitating Claude 3.7 sonnet thinking model release (current version)
• 🔄 Mathematics problem-solving specialized version
• 🔄 Domain-specific ultra-lightweight models for law, medicine, science, etc.
• 🔄 multilingual support with very small models(1K, 10k, 100K, etc parameters)
• 🔄 fine-tuning of models with very large maximum tokens but small parameter sizes

---