Downloads · 30 days
14
10% of all-time downloads
AmanPriyanshu/gpt-oss-18.5b-specialized-science-pruned-moe-only-28-experts
gpt-oss-18.5b-specialized-science-pruned-moe-only-28-experts is a text generation model from AmanPriyanshu. Use it when you need the model to write or continue text. The card lists the license as apache-2.0.
Project: https://amanpriyanshu.github.io/GPT-OSS-MoE-ExpertFingerprinting/
Downloads · 30 days
14
10% of all-time downloads
All-time downloads
135
Public
Parameters
18.5B
37.1 GB on disk
Likes
1
Public
Click a slice to open those files.
.safetensors37 GB · 100%
From the Hugging Face model README
Project: https://amanpriyanshu.github.io/GPT-OSS-MoE-ExpertFingerprinting/
<div align="center">This is a pruned variant of OpenAI's GPT-OSS-20B model, reduced to 28 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 science 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.
| Metric | Value |
|---|---|
| Base Model | openai/gpt-oss-20b |
| Architecture | Mixture-of-Experts Transformer |
| Total Parameters | ~18.5B (pruned from 21B) |
| Original Experts per Layer | 32 |
| Pruned Experts per Layer | 28 |
| 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 | Science |
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:
Note: Performance may vary depending on how well the pruned experts match your specific use case.
This science-specialized model leverages experts that showed high activation patterns during scientific reasoning tasks from GPQA (physics, chemistry, biology) and MMLU science domains. These experts demonstrate superior performance on complex scientific problem-solving and technical knowledge recall.
The expert selection process utilized our comprehensive analysis of router activation patterns across multiple evaluation benchmarks:
By identifying experts that consistently activated for science tasks, we created this specialized model that maintains domain expertise while significantly reducing computational requirements from 32 to 28 experts per layer.
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.
Our approach involves:
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.
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
# Load the specialized model on CPU
model = AutoModelForCausalLM.from_pretrained(
"AmanPriyanshu/gpt-oss-18.5b-specialized-science-pruned-moe-only-28-experts",
torch_dtype=torch.bfloat16,
device_map="cpu",
trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-18.5b-specialized-science-pruned-moe-only-28-experts")
# Generate with the model
messages = [
{"role": "user", "content": "Explain the process of photosynthesis in plants."}
]
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)
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-18.5b-specialized-science-pruned-moe-only-28-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-18.5b-specialized-science-pruned-moe-only-28-experts")
# Generate with the model
messages = [
{"role": "user", "content": "Explain the process of photosynthesis in plants."}
]
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)
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
The following examples demonstrate the model's performance on science tasks:
Question: Explain the process of photosynthesis in plants.
Response:
<|channel|>analysis<|message|>Need explain process.<|end|><|start|>assistant<|channel|>final<|message|>**Photosynthesis** is the process by which green plants, algae, and some bacteria convert light energy from the sun into chemical energy stored in organic molecules. In plants, this occurs mainly in the chloroplasts of leaf cells. The overall reaction can be written as:
\[
6\,\text{CO}_2 + 6\,\text{H}_2\text{O} + \text{light energy} \;\longrightarrow\; C_6H_{12}O_6 + 6\,\text{O}_2
\]
In other words, six molecules of carbon dioxide and six molecules of water, using light energy, are turned into one molecule of glucose (a simple sugar) and six molecules of oxygen.
The process is divided into two main stages:
| Stage | Key events | Main products | Where it happens |
|-------|------------|---------------|------------------|
| **Light‑dependent reactions (Photosystems I & II)** | 1. Light photons are absorbed by chlorophyll and accessory pigments. <br>2. Excited electrons travel through the electron transport chain, creating a proton gradient. <br>3. ATP synthase uses the gradient to produce ATP. <br>4. NADP⁺ is reduced to NADPH. | **ATP** (energy currency) and **NADPH** (reducing power) | Thylakoid membranes (inside the chloroplast) |
| **Calvin‑Benson cycle (Light‑independent reactions)** | 1. CO₂ is fixed into 3‑phosphoglycerate (3‑PGA) by the enzyme Rubisco. <br>2. 3‑PGA is phosphorylated by ATP and reduced by NADPH to glyceraldehyde‑3‑phosphate (G3P). <br>3. Some G3P molecules exit the cycle to form glucose and other carbohydrates. <br>4. Remaining G3P is used to regenerate ribulose‑1,5‑bisphosphate (RuBP), the CO₂ acceptor. | **Glucose** (and other sugars), **O₂** (released from water) | Stroma (the fluid surrounding the thylakoids) |
### Detailed Steps
1. **Light Absorption**
- Chlorophyll a and b, carotenoids, and other pigments absorb photons.
- The energy excites electrons in chlorophyll
Question: What causes the greenhouse effect and how does it work?
Response:
<|channel|>analysis<|message|>Need explain greenhouse effect, greenhouse gases, absorption, re-emission, energy balance.<|end|><|start|>assistant<|channel|>final<|message|>### The Greenhouse Effect – What It Is and How It Works
| **Step** | **What Happens** | **Why It Matters** |
|----------|------------------|--------------------|
| 1. Solar radiation reaches Earth | Sun emits a spectrum of electromagnetic waves, mainly visible light and a bit of infrared (IR). | This is the primary energy source for Earth’s climate. |
| 2. Surface absorbs and warms | The Earth’s surface (land, oceans, ice) absorbs most of the incoming solar energy and heats up. | The surface temperature rises to a level that balances the incoming energy. |
| 3. Surface re‑radiates energy | The warmed surface emits energy back into space, but this time it is mostly in the long‑wave (infrared) part of the spectrum. | Infrared radiation is the key to the greenhouse effect. |
| 4. Greenhouse gases absorb IR | Certain gases in the atmosphere—chiefly water vapor (H₂O), carbon dioxide (CO₂), methane (CH₄), nitrous oxide (N₂O), and others—have molecular vibrations that resonate with IR wavelengths. They absorb some of the outgoing IR. | These gases are “transparent” to visible light but “opaque” to IR, so they trap heat. |
| 5. Re‑emission in all directions | After absorbing IR, the greenhouse gas molecules re‑emit radiation in all directions, including back toward the surface. | This back‑radiation warms the surface further, creating a positive feedback loop. |
| 6. Energy balance shifts | The surface now receives more energy (solar + back‑radiation) than it loses to space, so it warms until a new equilibrium is reached. | The planet’s average temperature rises compared to a world without greenhouse gases. |
---
## 1. What Causes the Greenhouse Effect?
| **Natural** | **Anthropogenic (Human‑Made)** |
|-------------|-------------------------------|
| **Water vapor** – the most abundant greenhouse gas, produced by evaporation. | **Carbon dioxide** – from burning fossil fuels, deforestation, cement production. |
| **Methane** – from wetlands, ruminants, rice paddies. | **Methane** – from agriculture, landfills, natural gas leaks. |
| **Nitrous oxide** – from soils, fertilizers.
Question: Describe the structure and function of DNA.
Response:
<|channel|>analysis<|message|>Need concise description.<|end|><|start|>assistant<|channel|>final<|message|>**DNA (Deoxyribonucleic Acid)** is the hereditary material that carries the genetic instructions used in the growth, development, functioning, and reproduction of all known living organisms and many viruses. Its structure and function can be understood in two complementary parts: the physical architecture of the molecule and the biological roles it performs.
---
## 1. Structural Overview
| Feature | Description | Key Details |
|---------|-------------|-------------|
| **Polymers** | DNA is a polymer of nucleotides. | Each nucleotide consists of a deoxyribose sugar, a phosphate group, and one of four nitrogenous bases. |
| **Bases** | Adenine (A), Thymine (T), Cytosine (C), Guanine (G). | A pairs with T via two hydrogen bonds; C pairs with G via three hydrogen bonds. |
| **Backbone** | 5′‑phosphate → 3′‑deoxyribose → 5′‑phosphate → … | The sugar‑phosphate backbone is antiparallel; one strand runs 5′→3′, the complementary strand 3′→5′. |
| **Double Helix** | Two strands wound around each other. | Classic B‑form helix: ~10.5 base pairs per turn, ~3.4 Å rise per base pair. |
| **Major/Minor Grooves** | Spaces between the two strands. | Provide sites for protein binding and enzymatic activity. |
| **Supercoiling** | DNA can be over‑ or under‑wound. | Maintains compactness in cells and facilitates processes like replication and transcription. |
| **Chromatin Packaging** | DNA wrapped around histone proteins. | Forms nucleosomes → 30‑nm fiber → higher‑order structures. |
---
## 2. Functional Roles
| Function | How DNA Performs It | Biological Significance |
|----------|---------------------|------------------------|
| **Genetic Information Storage** | Sequence of bases encodes genes. | Determines phenotype, cellular function, and inheritance. |
| **Replication** | Semi‑conservative duplication via DNA polymerases. | Produces identical copies for cell division. |
| **Transcription** | RNA polymerase reads DNA template → mRNA. | Converts genetic code into messenger RNA for protein synthesis. |
| **Translation (via mRNA)** | Ribosomes read mRNA cod
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}
}