You can deploy this model and start earning money today!

AmanPriyanshu/
gpt-oss-14.3b-specialized-health_or_medicine-pruned-moe-only-21-experts

Name: gpt-oss-14.3b-specialized-health_or_medicine-pruned-moe-only-21-experts
Availability: InStock

Views

No views yet

Health Or Medicine GPT-OSS Model (21 Experts)

Project: https://amanpriyanshu.github.io/GPT-OSS-MoE-ExpertFingerprinting/

👥 Follow the Authors

Aman Priyanshu

Supriti Vijay

Introduction

This is a pruned variant of OpenAI's GPT-OSS-20B model, reduced to 21 experts per layer based on activation patterns from the AmanPriyanshu/GPT-OSS-20B MoE Expert Activations dataset. We analyzed router decisions across evaluation benchmarks to identify and retain experts most relevant for health or medicine tasks.

⚠️ Experimental Model: This is an experimental pruned model that may not work well - check the examples below to see if the outputs meet your needs before use.

This pruning approach reduces the model size while attempting to preserve performance on the target domain.

Model Architecture & Statistics

Metric	Value
Base Model	openai/gpt-oss-20b
Architecture	Mixture-of-Experts Transformer
Total Parameters	~14.3B (pruned from 21B)
Original Experts per Layer	32
Pruned Experts per Layer	21
Layers	24
Top-k Routing	4
Context Length	128K tokens
Attention Heads	64 (Query), 8 (Key-Value)
Residual Dimension	2880
Attention Pattern	Alternating dense & sliding window (128 tokens)
Positional Encoding	RoPE (Rotary Position Embedding)
Normalization	RMSNorm
Precision	BF16
License	Apache 2.0
Specialization	Health Or Medicine

Pruning Methodology

What is Expert Pruning?

Mixture-of-Experts models contain multiple specialized sub-networks (experts) per layer. During inference, only a subset of experts are activated for each token. Expert pruning involves:

Analyzing Usage Patterns: Tracking which experts activate most frequently for specific tasks
Removing Underutilized Experts: Discarding experts with low activation rates for the target domain
Preserving Router Functionality: Maintaining the routing mechanism with fewer available experts

Our Approach

Data-Driven Selection: Used activation patterns from health or medicine evaluation tasks
Systematic Reduction: Reduced from 32 to 21 experts per layer
No Retraining: Direct removal without additional training steps

Performance & Applications

Pruning Benefits

Smaller Memory Footprint: 65.6% of original expert parameters
Reduced Computational Load: Fewer routing decisions during inference
Focused Capabilities: Retains experts relevant to health or medicine tasks

Use Cases

Speculative Decoding: Draft model for full GPT-OSS-20B
Resource-Constrained Deployment: Edge devices, mobile applications
Research: Study expert specialization in MoE models
Fine-tuning: Smaller base model for domain adaptation

Note: Performance may vary depending on how well the pruned experts match your specific use case.

Motivation & Expert Selection

This medical domain model incorporates experts that activated highly during health and medical tasks from MMLU medical subjects. These experts specialize in clinical knowledge, anatomy, medical procedures, and health-related reasoning.

The expert selection process utilized our comprehensive analysis of router activation patterns across multiple evaluation benchmarks:

GPQA: Graduate-level questions in physics, chemistry, biology (Diamond & Expert subsets)
MMLU/MMLU-Pro: Comprehensive knowledge across 57+ subjects including science, medicine, law
SORRY-Bench: Safety evaluation across harmful content categories
Tulu3: Persona-driven instruction following with verifiable constraints
Polyglot-or-Not: Multilingual factual completion tasks

By identifying experts that consistently activated for health or medicine tasks, we created this specialized model that maintains domain expertise while significantly reducing computational requirements from 32 to 21 experts per layer.

Dataset & Analysis Foundation

This model is based on analysis from the GPT-OSS-20B MoE Expert Activations dataset available at: 🔗 https://huggingface.co/datasets/AmanPriyanshu/GPT-OSS-20B-MoE-expert-activations

The dataset contains router activation patterns from OpenAI's GPT-OSS-20B model across diverse evaluation benchmarks, enabling the creation of these domain-optimized models through systematic expert pruning.

Pruning Methodology

Our approach involves:

Activation Analysis: Comprehensive evaluation of expert usage patterns across domain-specific tasks
Expert Ranking: Identification of the most frequently activated experts for target domains
Systematic Pruning: Reduction from 32 to 21 experts while preserving router functionality
Quality Validation: Testing to ensure maintained performance on target tasks

This is a direct pruning approach - no additional training was performed. The model inherits all capabilities from the original GPT-OSS-20B with focused expert selection.

Usage

CPU Inference

python

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Load the specialized model on CPU
model = AutoModelForCausalLM.from_pretrained(
    "AmanPriyanshu/gpt-oss-14.3b-specialized-health_or_medicine-pruned-moe-only-21-experts", 
    torch_dtype=torch.bfloat16, 
    device_map="cpu", 
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-14.3b-specialized-health_or_medicine-pruned-moe-only-21-experts")

# Generate with the model
messages = [
    {"role": "user", "content": "What are the main functions of the human heart?"}
]

inputs = tokenizer.apply_chat_template(
    messages, 
    add_generation_prompt=True, 
    return_tensors="pt", 
    return_dict=True,
    reasoning_effort="medium"
)

# Ensure inputs are on the same device as model
inputs = {k: v.to(model.device) for k, v in inputs.items()}

outputs = model.generate(
    **inputs, 
    max_new_tokens=512,
    do_sample=True,
    temperature=0.1,
    top_p=0.9,
    pad_token_id=tokenizer.eos_token_id,
    eos_token_id=tokenizer.eos_token_id
)

# Decode only the generated part
input_length = inputs['input_ids'].shape[1]
response_tokens = outputs[0][input_length:]
response = tokenizer.decode(response_tokens, skip_special_tokens=True)
print(response)

Apple Silicon (MPS) Inference

python

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Check MPS availability and load model
device = "mps" if torch.backends.mps.is_available() else "cpu"

model = AutoModelForCausalLM.from_pretrained(
    "AmanPriyanshu/gpt-oss-14.3b-specialized-health_or_medicine-pruned-moe-only-21-experts", 
    torch_dtype=torch.float16,  # Better MPS compatibility
    device_map=device, 
    trust_remote_code=True,
    low_cpu_mem_usage=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-14.3b-specialized-health_or_medicine-pruned-moe-only-21-experts")

# Generate with the model
messages = [
    {"role": "user", "content": "What are the main functions of the human heart?"}
]

inputs = tokenizer.apply_chat_template(
    messages, 
    add_generation_prompt=True, 
    return_tensors="pt", 
    return_dict=True,
    reasoning_effort="medium"
)

# Move inputs to model device
inputs = {k: v.to(model.device) if isinstance(v, torch.Tensor) else v for k, v in inputs.items()}

# Use torch.no_grad for MPS stability
with torch.no_grad():
    outputs = model.generate(
        **inputs, 
        max_new_tokens=512,
        do_sample=True,
        temperature=0.1,
        top_p=0.9,
        pad_token_id=tokenizer.eos_token_id,
        eos_token_id=tokenizer.eos_token_id,
        use_cache=True
    )

# Decode only the generated part
input_length = inputs['input_ids'].shape[1]
response_tokens = outputs[0][input_length:]
response = tokenizer.decode(response_tokens, skip_special_tokens=True)
print(response)

GPU Inference

For CUDA GPU inference, the models work with minimal modifications from the CPU example above. Simply change:

python

device_map="auto"  # Will automatically use GPU if available
torch_dtype=torch.bfloat16  # or torch.float16

Model Examples

The following examples demonstrate the model's performance on health or medicine tasks:

Example 1

Question: What are the main functions of the human heart?

Response:

<|channel|>analysis<|message|>Need concise answer.<|end|><|start|>assistant<|channel|>final<|message|>The human heart performs three core functions that keep the body alive and functioning:

| Function | What it does | How it works |
|----------|--------------|--------------|
| **Pumping** | Moves blood throughout the body | The left ventricle contracts to push oxygen‑rich blood into the aorta; the right ventricle pumps de‑oxygenated blood into the pulmonary artery. |
| **Circulation** | Delivers oxygen, nutrients, hormones, and removes waste | Blood travels through arteries → capillaries → veins, exchanging substances with tissues. |
| **Regulation** | Maintains blood pressure and flow | The heart’s intrinsic pacemaker (SA node) sets rhythm; autonomic nerves and hormones adjust rate and force of contraction to meet the body’s needs. |

Together, these functions keep tissues supplied with oxygen and nutrients while removing metabolic waste, enabling the body to function and respond to changing demands.<|return|>

Example 2

Question: Explain the difference between bacteria and viruses.

Response:

<|channel|>analysis<|message|>Need concise explanation.<|end|><|start|>assistant<|channel|>final<|message|>**Bacteria vs. Viruses – the key differences**

| Feature | Bacteria | Viruses |
|--------|----------|--------|
| **Size** | Microscopic, 0.5–5 µm (visible under a microscope) | 20–200 nm (too small for a light microscope) |
| **Structure** | Complex, cell‑like structure: cell wall, membrane, cytoplasm, organelles (e.g., ribosomes) | Very simple: a nucleic acid genome (DNA or RNA) surrounded by a protein coat (capsid); some have a lipid envelope |
| **Life Cycle** | Independent life form; can grow, divide, and reproduce on their own (metabolism, energy production) | Not a living organism on its own; must infect a host cell and hijack its machinery to replicate |
| **Reproduction** | Divide by binary (or sometimes more complex) cell division (mitosis‑like) | Replicate by copying their genome inside a host cell, then assembling new virions |
| **Genetic Material** | DNA (usually double‑stranded) | DNA or RNA (single‑ or double‑stranded) |
| **Infection Mechanism** | Can colonize or invade tissues, produce toxins, or cause inflammation; often spread by contact, food, water, vectors | Attach to specific host receptors, enter cells, use host enzymes to replicate; often cause cell lysis or dormancy |
| **Treatment** | Antibiotics (target cell wall, protein synthesis, DNA replication, etc.) | Antiviral drugs (target viral enzymes or replication steps); vaccines; no antibiotics work |
| **Evolutionary Complexity** | More complex genomes, can exchange genes via plasmids, conjugation, etc. | Simpler genomes; evolve mainly by mutation and recombination |
| **Examples** | *Staphylococcus aureus*, *E. coli*, *Bacillus subtilis* | Influenza virus, HIV, SARS‑CoV‑2, HPV |

**Bottom line:**  
Bacteria are true, independent microorganisms that can live and reproduce on their own, whereas viruses are essentially “parasite‑like” particles that need a host cell to make copies of themselves. This fundamental difference shapes how they cause disease, how we diagnose them, and how we treat or prevent infections.<|return|>

Example 3

Question: What are the symptoms and causes of diabetes?

Response:

<|channel|>analysis<|message|>Need concise answer.<|end|><|start|>assistant<|channel|>final<|message|>**Diabetes mellitus** is a chronic metabolic disorder in which the body cannot regulate blood glucose properly.  The two main types—Type 1 and Type 2—share many symptoms but differ in underlying causes.

---

## 1. Symptoms (common to both types)

| Symptom | Typical Onset | Why it Happens |
|---------|---------------|----------------|
| **Polyuria** (excessive urination) | Often first sign | High blood glucose spills into urine, pulling water with it (osmotic diuresis). |
| **Polydipsia** (excessive thirst) | Usually follows polyuria | Loss of fluid through urine leads to dehydration; kidneys signal thirst. |
| **Polyphagia** (increased hunger) | Common in Type 1; variable in Type 2 | Cells can’t get glucose → body signals “need more food.” |
| **Weight loss** | More pronounced in Type 1 | Body breaks down fat and muscle for energy when glucose can’t be used. |
| **Fatigue / weakness** | Early symptom | Energy deficit from impaired glucose utilization. |
| **Blurred vision** | Can appear early | Hyperglycemia causes fluid shifts in the lens, altering its shape. |
| **Slow‑healing cuts or infections** | Later symptom | High glucose impairs immune function and tissue repair. |
| **Numbness or tingling in extremities** | Later symptom | Chronic hyperglycemia damages nerves (diabetic neuropathy). |
| **Frequent infections (e.g., urinary, skin, oral)** | Later symptom | Impaired immune response. |

> **Note:** In Type 2 diabetes, many people are asymptomatic for years; early signs may be subtle (e.g., mild fatigue, occasional blurred vision).

---

## 2. Causes

| Type | Primary Cause | Key Risk Factors |
|------|---------------|-----------------|
| **Type 1** | Autoimmune destruction of pancreatic β‑cells → insulin deficiency | Genetic predisposition, viral infections (e.g., enteroviruses), environmental triggers, early childhood onset |
| **Type 2** | Insulin resistance + relative insulin deficiency | Obesity, sedentary lifestyle, poor diet, family history, age ≥45, certain ethnicities (African‑American, Hispanic, Native American, Asian), gestational diabetes history |
| **Gestational** | Hormonal changes during pregnancy → insulin resistance

Citation

If you use this model in your research, please cite:

bibtex

@misc{priyanshu2025gptoss,
  title={{GPT-OSS MoE Expert Fingerprinting: Analyzing Expert Activation Patterns in Mixture of Experts Models}},
  author={Priyanshu, Aman and Vijay, Supriti},
  year={2025},
  howpublished={\url{https://amanpriyanshu.github.io/GPT-OSS-MoE-ExpertFingerprinting/}},
  note={Interactive analysis tool for expert activation patterns in MoE architectures}
}

References & Resources

Original Model: OpenAI GPT-OSS Model Card
Model Hub: GPT-OSS-20B on Hugging Face
Expert Analysis Dataset: GPT-OSS-20B MoE Expert Activations
Project Page: GPT-OSS MoE Expert Fingerprinting
GitHub Repository: OpenAI GPT-OSS