You can deploy this model and start earning money today!

AmanPriyanshu/
gpt-oss-14.3b-specialized-law-pruned-moe-only-21-experts

Name: gpt-oss-14.3b-specialized-law-pruned-moe-only-21-experts
Availability: InStock

Views

No views yet

Law GPT-OSS Model (21 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 21 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 law 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	~14.3B (pruned from 21B)
Original Experts per Layer	32
Pruned Experts per Layer	21
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	Law

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 law evaluation tasks
Systematic Reduction: Reduced from 32 to 21 experts per layer
No Retraining: Direct removal without additional training steps

Performance & Applications

Pruning Benefits

Smaller Memory Footprint: 65.6% of original expert parameters
Reduced Computational Load: Fewer routing decisions during inference
Focused Capabilities: Retains experts relevant to law 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 legal domain model employs experts that demonstrated expertise during law-related tasks from MMLU legal subjects. These experts excel at legal reasoning, jurisprudence, and understanding of legal frameworks and procedures.

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 law tasks, we created this specialized model that maintains domain expertise while significantly reducing computational requirements from 32 to 21 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 21 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-14.3b-specialized-law-pruned-moe-only-21-experts", 
    torch_dtype=torch.bfloat16, 
    device_map="cpu", 
    trust_remote_code=True
)
tokenizer = AutoTokenizer.from_pretrained("AmanPriyanshu/gpt-oss-14.3b-specialized-law-pruned-moe-only-21-experts")

# Generate with the model
messages = [
    {"role": "user", "content": "What is the difference between civil and criminal law?"}
]

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-14.3b-specialized-law-pruned-moe-only-21-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-14.3b-specialized-law-pruned-moe-only-21-experts")

# Generate with the model
messages = [
    {"role": "user", "content": "What is the difference between civil and criminal law?"}
]

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 law tasks:

Example 1

Question: What is the difference between civil and criminal law?

Response:

<|channel|>analysis<|message|>Need concise explanation.<|end|><|start|>assistant<|channel|>final<|message|>**Civil law** and **criminal law** are the two main branches of the legal system, and they serve very different purposes:

| Feature | Civil Law | Criminal Law |
|---------|-----------|--------------|
| **Purpose** | Resolve disputes between private parties (individuals, businesses, organizations). | Protect society by punishing conduct that is deemed harmful or dangerous to the public. |
| **Parties** | Plaintiff (the person who brings the claim) vs. Defendant (the person being sued). | State (or government) vs. Defendant (the accused). |
| **Burden of Proof** | “Pre‑ponderance of evidence” – more likely than unlikely. | “Beyond a reasonable doubt” – the highest standard. |
| **Typical Cases** | Contract disputes, property issues, torts (e.g., negligence), family matters (divorce, custody), estate/ probate. | Theft, assault, fraud, homicide, drug offenses, traffic violations. |
| **Remedies** | Compensatory damages (money), injunctions (orders to do or stop doing something), specific performance (e.g., delivering a contract). | Penalties such as fines, imprisonment, probation, community service, or other sanctions. |
| **Procedure** | Usually civil courts; parties may choose to settle or go to trial. | Criminal courts; the prosecution must prove guilt; defendants have rights such as the right to counsel, to remain silent, and to a jury trial. |
| **Outcome** | Restitution or correction of the wrong; no criminal punishment. | Punishment or correction of the offender; may also include restitution to victims. |

In short: **civil law** is about *compensation and rights* between private parties, while **criminal law** is about *punishment and deterrence* for offenses that harm society.<|return|>

Example 2

Question: Explain the concept of due process in legal systems.

Response:

<|channel|>analysis<|message|>Need explain due process.<|end|><|start|>assistant<|channel|>final<|message|>**Due Process (or Due Process) – A Fundamental Legal Principle**

---

### 1. What is Due Process?

**Due process** is the set of procedural safeguards that a legal system must provide to ensure that decisions affecting a person’s rights, property, or liberty are made fairly, transparently, and with an opportunity for the affected party to be heard. It is the procedural counterpart to the substantive rights protected by a constitution or law.

- **“Due”** means *proper, fair, and lawful*.
- **Process** refers to the *sequence of steps* (notice, hearing, decision, appeal, etc.) that a case must follow.

---

### 2. Core Elements of Due Process

| Element | What it Means | Why It Matters |
|---------|----------------|----------------|
| **Notice** | The person must be informed of the claim, charge, or action against them, and of the time, place, and manner of the proceeding. | Prevents surprise and allows preparation. |
| **Opportunity to Respond** | The person must be given a chance to present evidence, argue, and cross‑examine witnesses. | Ensures that the decision is based on a balanced view. |
| **Impartial Decision‑Maker** | The judge, magistrate, or tribunal must be unbiased and free from conflicts of interest. | Guarantees fairness and public confidence. |
| **Reasoned Decision** | The decision must be based on facts, evidence, and applicable law, and the reasoning must be written and accessible. | Allows review, appeals, and accountability. |
| **Right to Appeal** | The person can challenge the decision in a higher court or tribunal. | Provides a safety net against errors or abuse. |
| **Timeliness** | Proceedings should be conducted within a reasonable time. | Prevents indefinite uncertainty and protects liberty. |
| **Equality Before the Law** | All persons receive the same procedural protections, regardless of status. | Upholds the principle of equal treatment. |

---

### 3. How Due Process Differs Across Legal Systems

| System | Typical Features | Variations |
|--------|-----------------|------------|
| **Common Law (e.g., U.S., U.K.)** | Adversarial system; strong emphasis on oral argument, cross‑examination, and written judgments. | In the U.S., the *Constrained Due Process* (e.g., in criminal cases) is stric

Example 3

Question: What are the basic rights protected by the First Amendment?

Response:

<|channel|>analysis<|message|>Need to answer: freedom of religion, speech, press, assembly, petition.<|end|><|start|>assistant<|channel|>final<|message|>The First Amendment to the United States Constitution protects several core civil liberties that are considered fundamental to a free society.  In plain terms, it guarantees:

| Right | What it means |
|-------|--------------|
| **Freedom of religion** | Individuals may practice any religion, or none, without government interference. |
| **Freedom of speech** | People can express ideas, opinions, and information—whether spoken, written, or symbolic—without censorship or punishment. |
| **Freedom of the press** | Newspapers, books, radio, television, and other media can publish news and viewpoints without government control. |
| **Freedom of assembly** | Citizens may gather peacefully for protests, rallies, or other collective activities. |
| **Right to petition the government** | Individuals can formally request the government to address grievances or change laws. |

These five protections together form the core of the First Amendment’s guarantee of individual liberty and democratic participation.<|return|>

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