Downloads · 30 days
15
11% of all-time downloads
AmanPriyanshu/gpt-oss-17.9b-specialized-science-pruned-moe-only-27-experts
gpt-oss-17.9b-specialized-science-pruned-moe-only-27-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
15
11% of all-time downloads
All-time downloads
134
Public
Parameters
17.9B
35.9 GB on disk
Likes
1
Public
Click a slice to open those files.
.safetensors35.9 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 27 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 | ~17.9B (pruned from 21B) |
| Original Experts per Layer | 32 |
| Pruned Experts per Layer | 27 |
| 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 27 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-17.9b-specialized-science-pruned-moe-only-27-experts",
torch_dtype=torch.bfloat16,
device_map="cpu",
trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-17.9b-specialized-science-pruned-moe-only-27-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-17.9b-specialized-science-pruned-moe-only-27-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-17.9b-specialized-science-pruned-moe-only-27-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 concise explanation.<|end|><|start|>assistant<|channel|>final<|message|>**Photosynthesis** is the process by which green plants, algae, and some bacteria convert light energy into chemical energy, producing glucose (a sugar) and oxygen from carbon dioxide and water. It occurs mainly in the chloroplasts of plant cells, specifically in the thylakoid membranes and the stroma.
---
## 1. Light‑Dependent Reactions (Thylakoid Membrane)
| Step | What Happens | Key Components |
|------|--------------|----------------|
| **Photon Capture** | Light (visible spectrum) hits chlorophyll‑a and accessory pigments. | Chlorophyll‑a, chlorophyll‑b, carotenoids |
| **Excitation & Electron Transport** | Excited electrons travel through the *photosynthetic electron transport chain* (P680 → plastoquinone → cytochrome b₆f → plastocyanin → Photosystem II). | Photosystem II (PSII), Photosystem I (PSI), plastoquinone, cytochrome b₆f, plastocyanin |
| **Water Splitting (Oxygen Evolution)** | PSII splits H₂O into O₂, protons (H⁺), and electrons. | Water‑splitting complex (Mn‑cluster) |
| **ATP Synthesis** | Proton gradient drives ATP synthase to produce ATP. | ATP synthase, proton gradient |
| **NADPH Formation** | Electrons reduce NADP⁺ to NADPH at PSI. | NADP⁺ reductase |
**Products of Light‑Dependent Reactions**
- 2 ATP (per NADPH produced)
- 2 NADPH
- ½ O₂ (released to atmosphere)
---
## 2. Light‑Independent Reactions (Calvin Cycle, Stroma)
| Step | What Happens | Key Enzymes |
|------|--------------|-------------|
| **Carbon Fixation** | CO₂ is attached to ribulose‑1,5‑bisphosphate (RuBP) by Rubisco, forming 3‑phosphoglycerate (3‑PGA). | Rubisco (ribulose‑1,5‑bisphosphate carboxylase/oxygenase) |
| **Reduction Phase** | 3‑PGA is phosphorylated by ATP and reduced by NADPH to glyceraldehyde‑3‑phosphate (G3P). | ATP‑dependent kinase
Question: What causes the greenhouse effect and how does it work?
Response:
<|channel|>analysis<|message|>Need explain greenhouse effect, causes, mechanism.<|end|><|start|>assistant<|channel|>final<|message|>### What Causes the Greenhouse Effect?
The greenhouse effect is a natural process that warms the Earth’s surface. It is caused by certain gases in the atmosphere—called **greenhouse gases (GHGs)**—that absorb and re‑emit infrared (IR) radiation. The main greenhouse gases are:
| Gas | Typical Atmospheric Concentration | Primary Source |
|-----|------------------------------------|----------------|
| **Water vapor (H₂O)** | 0.1 – 4 % (variable) | Evaporation from oceans, lakes, soil, plants |
| **Carbon dioxide (CO₂)** | ~0.04 % (400 ppm) | Fossil‑fuel combustion, deforestation, industrial processes |
| **Methane (CH₄)** | ~0.0002 % (1.8 ppm) | Agriculture (especially ruminants), livestock, rice paddies, landfills, natural gas leaks |
| **Nitrous oxide (N₂O)** | ~0.00003 % (0.3 ppm) | Agricultural fertilizers, industrial processes, combustion |
| **Ozone (O₃)** | ~0.00001 % (0.1 ppm) | Photochemical reactions in the troposphere |
| **Fluorinated gases (HFCs, PFCs, SF₆, etc.)** | Trace | Refrigeration, electronics, industrial processes |
These gases are **transparent to visible sunlight** (the short‑wave radiation that comes from the Sun) but **absorb long‑wave infrared radiation** emitted by the Earth’s surface. That absorption and subsequent re‑emission is what keeps the planet warm.
---
### How the Greenhouse Effect Works (Step‑by‑Step)
1. **Solar Radiation Reaches Earth**
- The Sun emits a spectrum of electromagnetic radiation.
- About 30 % of the incoming solar energy is reflected back to space by clouds, ice, and the atmosphere (the *albedo*).
- The remaining ~70 % is absorbed by the Earth’s surface (land, oceans, vegetation).
2. **Earth’s Surface Heats Up**
- The absorbed solar energy warms the surface, raising its temperature.
- A warm surface emits thermal radiation (infrared, long‑wave radiation) according to the Stefan–Boltzmann law:
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 ways: **molecular architecture** and **biological role**.
---
## 1. Molecular Structure
| Feature | Description | Biological Significance |
|---------|-------------|------------------------|
| **Backbone** | Alternating sugar (deoxyribose) and phosphate groups. | Provides structural stability and a negatively charged scaffold that attracts proteins and ions. |
| **Bases** | Four nitrogenous bases: Adenine (A), Thymine (T), Cytosine (C), Guanine (G). | Encode genetic information; base‑pairing (A‑T, C‑G) ensures faithful replication. |
| **Double‑helix** | Two complementary strands wound around each other. | Protects genetic material, allows efficient packing in the nucleus, and facilitates replication and transcription. |
| **Major & Minor grooves** | Unequal spaces between the strands. | Sites for protein binding (e.g., transcription factors, polymerases). |
| **Antiparallel orientation** | One strand runs 5’→3’, the other 3’→5’. | Essential for the directionality of replication and transcription enzymes. |
| **Supercoiling & nucleosomes** | DNA is further compacted by winding around histone proteins. | Enables regulation of gene accessibility and efficient packaging in chromosomes. |
---
## 2. Functional Roles
| Function | Mechanism | Key Players |
|----------|-----------|-------------|
| **Genetic Information Storage** | Sequence of bases encodes proteins and regulatory RNAs. | Genes, introns, exons, promoters. |
| **Replication** | Semi‑conservative duplication: each new double helix contains one parental strand. | DNA polymerases, helicases, primases, ligases. |
| **Transcription** | RNA polymerase reads one DNA strand to synthesize messenger RNA (mRNA). | RNA polymerase II (eukaryotes), sigma factors (bacteria). |
| **Translation (Protein Synthesis)** | mRNA is translated by ribosomes into polypeptide chains. | tRNA, ribosomal RNA (rRNA), aminoacyl‑tRNA synthetases. |
| **Repair
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}
}