Downloads · 30 days
11
9% of all-time downloads
AmanPriyanshu/gpt-oss-6.0b-specialized-instruction_following-pruned-moe-only-7-experts
gpt-oss-6.0b-specialized-instruction_following-pruned-moe-only-7-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
11
9% of all-time downloads
All-time downloads
124
Public
Parameters
6B
12 GB on disk
Likes
1
Public
Click a slice to open those files.
.safetensors12 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 7 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 instruction following 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 | ~6.0B (pruned from 21B) |
| Original Experts per Layer | 32 |
| Pruned Experts per Layer | 7 |
| 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 | Instruction Following |
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 instruction-following model leverages experts that excelled at constraint satisfaction tasks from Tulu3 Persona Instruction Following dataset. These experts specialize in precise adherence to user specifications and formatting requirements.
The expert selection process utilized our comprehensive analysis of router activation patterns across multiple evaluation benchmarks:
By identifying experts that consistently activated for instruction following tasks, we created this specialized model that maintains domain expertise while significantly reducing computational requirements from 32 to 7 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-6.0b-specialized-instruction_following-pruned-moe-only-7-experts",
torch_dtype=torch.bfloat16,
device_map="cpu",
trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-6.0b-specialized-instruction_following-pruned-moe-only-7-experts")
# Generate with the model
messages = [
{"role": "user", "content": "Write a formal email to a professor requesting a meeting, including: subject line, greeting, purpose, proposed times, and professional closing."}
]
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-6.0b-specialized-instruction_following-pruned-moe-only-7-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-6.0b-specialized-instruction_following-pruned-moe-only-7-experts")
# Generate with the model
messages = [
{"role": "user", "content": "Write a formal email to a professor requesting a meeting, including: subject line, greeting, purpose, proposed times, and professional closing."}
]
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 instruction following tasks:
Question: Write a formal email to a professor requesting a meeting, including: subject line, greeting, purpose, proposed times, and professional closing.
Response:
<|channel|>analysis<|message|>We need to produce a formal email to a professor requesting a meeting. Must include subject line, greeting, purpose, proposed times, and professional closing. So we need to write a formal email. The user wants a formal email. We'll produce a subject line, greeting, purpose, proposed times, professional closing. We'll produce a professional email. We'll include subject line: "Request for Meeting" or "Meeting Request". We'll greet "Dear Professor [Name]" or "Dear Professor Dr. X". We'll state purpose: "I would like schedule a meeting to discuss...". We'll propose times: "Monday 2-30, Tuesday 3-30, Wednesday 4-30" or something. We'll close: "Sincerely" or "Kind regards" etc. We'll produce a professional closing: "Sincerely, [Name]" etc. We'll produce a formal email. We'll include subject line. We'll produce a professional email. We'll produce a final answer.<|end|><|start|>assistant<|channel|>final<|message|>**Subject:** Request for a Meeting on Project Progress
Dear Professor Dr. Smith,
I hope you are doing well. I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of my thesis and the next steps for the upcoming conference.
I would like schedule a meeting to discuss the progress of
Question: Create a numbered list of steps for making scrambled eggs, using exactly 6 steps.
Response:
<|channel|>analysis<|message|>We need to produce a numbered list of steps for making scrambled eggs, using exactly 6 steps. So we need 6 steps. Each step numbered 1- 1- 2- etc. Exactly 6 steps. So we need 6 steps. Provide a numbered list. Ensure exactly 6 steps. No more steps. So 1- 2- 3- 4- 5- 6. Provide concise.<|end|><|start|>assistant<|channel|>final<|message|>1. Heat a small pot or kettle to reach the right temperature for the eggs.
2. Crack the eggs into a small bowl or a small container.
3. Stir the eggs in a small bowl or a small container.
4. Flip the eggs into a small bowl or a small container.
5. Stir the eggs in a small bowl or a small container.
6. Flip the eggs into a small bowl or a small container.<|return|>
Question: Summarize the water cycle in exactly 50 words, no more, no less.
Response:
<|channel|>analysis<|message|>We need to produce exactly 50 words. Count words. Must be 50 words. Let's craft: "Water begins in the atmosphere, condurates, travels to the surface, forms droplets, mixes with other elements, evaporates, returns to the atmosphere, and repeats the cycle, sustaining life and ecosystems." Count words: Water(1) begins(2) in(3) the(4) atmosphere,(5) condurates,(6) travels(7) to(8) the(9) surface,(10) forms(11) droplets,(12) mixes(13) with(14) other(15) elements,(16) evaporates,(17) returns(18) to(19) the(20) atmosphere,(21) and(22) repeats(23) the(24) cycle,(27) sustaining(28) life(29) and(30) ecosystems(31). That's 29 words. Need 50 words. Let's craft: "Water originates in the atmosphere, condurates, travels to the surface, forms droplets, mixes with other elements, evaporates, returns to the atmosphere, and repeats the cycle, sustaining life and ecosystems." Count: Water(1) originates(2) in(3) the(4) atmosphere,(5) condurates,(6) travels(7) to(8) the(9) surface,(10) forms(11) droplets,(12) mixes(13) with(14) other(15) elements,(16) evaporates,(17) returns(18) to(19) the(20) atmosphere,(21) and(22) repeats(23) the(24) cycle,(27) sustaining(28) life(29) and(30) ecosystems(31). That's 31 words. Need 50. Let's add more: "Water originates in the atmosphere, condurates, travels to the surface, forms droplets, mixes with other elements, evaporates, returns to the atmosphere, and repeats the cycle, sustaining life and ecosystems, providing nourishment, energy, and sustaining countless species." Count: Water(1) originates(2) in(3) the(4) atmosphere,(5) condurates,(6) travels(7) to(8) the(9) surface,(10) forms(11) droplets,(12) mixes(13) with(14) other(15) elements,(16) evaporates,(17) returns(18) to(19) the(
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}
}