gpt-oss-4.2b-specialized-math-pruned-moe-only-4-experts

Math GPT-OSS Model (4 Experts)

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

<div align="center">

👥 Follow the Authors

Aman Priyanshu

Supriti Vijay

</div>

Introduction

This is a pruned variant of OpenAI's GPT-OSS-20B model, reduced to 4 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 math 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

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:

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

Our Approach

• Data-Driven Selection: Used activation patterns from math evaluation tasks
• Systematic Reduction: Reduced from 32 to 4 experts per layer
• No Retraining: Direct removal without additional training steps

Performance & Applications

Pruning Benefits

• Smaller Memory Footprint: 12.5% of original expert parameters
• Reduced Computational Load: Fewer routing decisions during inference
• Focused Capabilities: Retains experts relevant to math 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 mathematics-focused model utilizes experts that exhibited strong performance on mathematical reasoning tasks from MMLU mathematics subjects and quantitative sections. These experts excel at mathematical computation, proof strategies, and logical 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 math tasks, we created this specialized model that maintains domain expertise while significantly reducing computational requirements from 32 to 4 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:

1. Activation Analysis: Comprehensive evaluation of expert usage patterns across domain-specific tasks
2. Expert Ranking: Identification of the most frequently activated experts for target domains
3. Systematic Pruning: Reduction from 32 to 4 experts while preserving router functionality
4. 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

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Load the specialized model on CPU
model = AutoModelForCausalLM.from_pretrained(
    "AmanPriyanshu/gpt-oss-4.2b-specialized-math-pruned-moe-only-4-experts", 
    torch_dtype=torch.bfloat16, 
    device_map="cpu", 
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-4.2b-specialized-math-pruned-moe-only-4-experts")

# Generate with the model
messages = [
    {"role": "user", "content": "Solve this equation: 2x + 5 = 17. Show your work step by step."}
]

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

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-4.2b-specialized-math-pruned-moe-only-4-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-4.2b-specialized-math-pruned-moe-only-4-experts")

# Generate with the model
messages = [
    {"role": "user", "content": "Solve this equation: 2x + 5 = 17. Show your work step by step."}
]

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:

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 math tasks:

Example 1

Question: Solve this equation: 2x + 5 = 17. Show your work step by step.

Response:

<|channel|>analysis<|message|>We need to solve the equation: 2x + 5 = 17. Show step. The solution: 2x + 5 = 17 => subtract 5: 2x = 12 => divide by 2: x = 6. So step: 2x + 5 = 17; subtract 5: 2x = 12; divide by 2: x = 6. So show.<|end|><|start|>assistant<|channel|>final<|message|>**Step 1**  
\(2x + 5 = 17\)

**Step 2**  
Subtract 5 from each equation:  
\[
2x + 5 - 5 = 17 - 5 \Rightrightarrow 2x = 12
\]

**Step 3**  
Divide the whole by 2:  
\[
\frac{2x}{2} = \frac{12}{2} \Rightrightarrow x = 6
\]

**Answer**  
\(x = 6\)

So you just worked out the solution.<|return|>

---

Example 2

Question: What is the Pythagorean theorem and how is it used?

Response:

<|channel|>analysis<|message|>The user asks: "What is the Pythay theorem and how is it used?" The question: "What is the Pythay theorem and how is it used?" The answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and how is it used?" The user wants a short answer: "What is the Pythay theorem and

---

Example 3

Question: Calculate the area of a circle with radius 7 meters.

Response:

<|channel|>analysis<|message|>We need to calculate the area of a circle with radius 7 meters. The area of a circle is given formula: area = π r^2. r = 7. So area = π * 7^2 = π * 49 = 49π. So area is 49π. In meters squared? Actually 49π is in units of square of meters. So area is 49π m^2. So answer: 49π m^2. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49π. So answer: 49π. So we need to compute the area of a circle with radius 7 meters. So area = π * 49 = 49

---

Citation

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

@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