You can deploy this model and start earning money today!

AmanPriyanshu/
gpt-oss-20.3b-specialized-health_or_medicine-pruned-moe-only-31-experts

Name: gpt-oss-20.3b-specialized-health_or_medicine-pruned-moe-only-31-experts
Availability: InStock

Views

No views yet

Health Or Medicine GPT-OSS Model (31 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 31 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	~20.3B (pruned from 21B)
Original Experts per Layer	32
Pruned Experts per Layer	31
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 31 experts per layer
No Retraining: Direct removal without additional training steps

Performance & Applications

Pruning Benefits

Smaller Memory Footprint: 96.9% 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 31 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 31 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-20.3b-specialized-health_or_medicine-pruned-moe-only-31-experts", 
    torch_dtype=torch.bfloat16, 
    device_map="cpu", 
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-20.3b-specialized-health_or_medicine-pruned-moe-only-31-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-20.3b-specialized-health_or_medicine-pruned-moe-only-31-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-20.3b-specialized-health_or_medicine-pruned-moe-only-31-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 to answer.<|end|><|start|>assistant<|channel|>final<|message|>The human heart is a muscular pump that performs several essential functions to keep the circulatory system working efficiently. Its main functions can be grouped into four key categories:

| Function | What it does | Why it matters |
|----------|--------------|----------------|
| **Pumping blood** | The heart contracts (systole) and relaxes (diastole) in a coordinated rhythm, propelling blood through the body. | Provides continuous circulation of oxygen, nutrients, hormones, and waste products. |
| **Maintaining blood pressure** | By generating pressure during systole, the heart keeps blood moving against resistance in arteries and capillaries. | Ensures that tissues receive adequate perfusion and that the blood reaches all parts of the body. |
| **Regulating blood flow** | The heart’s rhythm and force of contraction can be adjusted by the autonomic nervous system and hormonal signals. | Allows the body to respond to changes in activity, stress, or metabolic demand (e.g., exercise, rest, fever). |
| **Supporting the circulatory system’s architecture** | The heart’s chambers (right/left atria and ventricles) and valves (tricuspid, pulmonary, mitral, aortic) create a one‑way flow of blood. | Prevents back‑flow, ensures efficient separation of oxygenated and deoxygenated blood, and maintains the overall integrity of the cardiovascular network. |

### How the heart accomplishes these tasks

1. **Electrical conduction system**  
   - **Sinoatrial (SA) node**: “pacemaker” that initiates the heartbeat.  
   - **Atrioventricular (AV) node, bundle of His, Purkinje fibers**: conduct the impulse to coordinate atrial and ventricular contraction.

2. **Mechanical contraction**  
   - **Myocardial fibers**: contract in a coordinated manner, first the atria, then the ventricles.  
   - **Valves**: open and close to direct blood flow and prevent back‑flow.

3. **Regulation**  
   - **Autonomic nervous system**: sympathetic stimulation increases heart rate and contractility; parasympathetic stimulation slows it.  
   - **Hormonal influences**: adrenaline, noradrenaline, thyroid hormones, and others modulate heart function.

4. **Interaction with the rest of the body**  
   - **Blood oxygenation**: oxygen‑rich blood is pumped

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 big differences**

| Feature | Bacteria | Viruses |
|---------|----------|---------|
| **Basic nature** | Living, single‑cell organisms (prokaryotes). | Non‑living infectious particles (no cellular structure). |
| **Size** | ~0.5–5 µm (visible under light microscope). | ~20–300 nm (visible only with electron microscope). |
| **Structure** | Cell wall, cytoplasm, ribosomes, DNA (circular or linear). | Protein coat (capsid) surrounding a nucleic acid core (DNA or RNA). |
| **Reproduction** | Binary fission (self‑replication). | Must hijack a host cell’s machinery to replicate. |
| **Metabolism** | Can metabolize nutrients, grow, and respond to stimuli. | No metabolism; rely entirely on host cell. |
| **Genetic material** | DNA (sometimes plasmids). | DNA or RNA, single or double stranded. |
| **Treatment** | Antibiotics (target cell wall, protein synthesis, etc.). | Antivirals (target viral enzymes or entry), vaccines. |
| **Role in disease** | Can cause infections (e.g., strep throat, urinary tract infection). | Cause infections (e.g., influenza, HIV, COVID‑19). |
| **Examples** | *Escherichia coli*, *Staphylococcus aureus*. | Influenza virus, HIV, SARS‑CoV‑2. |

**Bottom line:**  
Bacteria are autonomous, living cells that can grow and reproduce on their own. Viruses are tiny, non‑living particles that need a host cell to replicate. This fundamental difference shapes how we diagnose, treat, and prevent infections caused by each.<|return|>

Example 3

Question: What are the symptoms and causes of diabetes?

Response:

<|channel|>analysis<|message|>Need to answer.<|end|><|start|>assistant<|channel|>final<|message|>**Diabetes mellitus** is a group of metabolic disorders characterized by chronic hyperglycemia (high blood sugar). The two most common types are type 1 and type 2, but there are other less common forms (gestational, MODY, etc.). Below is a concise overview of the main symptoms, causes, and some additional details.

---

## 1. Symptoms

| Symptom | Typical Onset | Why It Happens |
|---------|---------------|----------------|
| **Polyuria** (frequent urination) | Early | Kidneys excrete excess glucose, pulling water with it. |
| **Polydipsia** (excessive thirst) | Early | Loss of fluid through urine leads to dehydration. |
| **Polyphagia** (increased hunger) | Early | Cells can’t use glucose, so the body signals for more food. |
| **Weight loss** (especially type 1) | Early | Body breaks down fat and muscle for energy. |
| **Fatigue / Weakness** | Early | Energy production is impaired. |
| **Blurred vision** | Early | Hyperglycemia causes fluid shifts in the lens. |
| **Slow‑healing cuts or infections** | Early | High glucose impairs immune function and circulation. |
| **Numbness or tingling in extremities** | Later | Chronic high glucose damages nerves (neuropathy). |
| **Darkened skin patches (acanthosis nigricans)** | Later | Often seen in insulin resistance (type 2). |
| **Recurrent urinary tract or vaginal infections** | Later | High glucose in urine provides a food source for bacteria. |
| **Night sweats, dizziness, or fainting** | Variable | Fluctuations in blood sugar or dehydration. |

> **Tip:** In type 1 diabetes, symptoms can appear rapidly over days to weeks. In type 2, they may develop gradually over months or years and can be subtle or absent until complications arise.

---

## 2. Causes

| Type | Primary Cause | Key Risk Factors |
|------|---------------|------------------|
| **Type 1 Diabetes** | Autoimmune destruction of pancreatic β‑cells → insulin deficiency | Genetic predisposition, viral infections (e.g., enteroviruses), environmental triggers, early childhood onset |
| **Type 2 Diabetes** | Insulin resistance + relative insulin deficiency | Obesity, sedentary lifestyle, poor diet

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