Instructions to use sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH")
model = AutoModelForCausalLM.from_pretrained("sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH

SGLang

How to use sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH with Docker Model Runner:
```
docker model run hf.co/sunblaze-ucb/Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH
```

Llama-3.2-3B-Instruct-Intuitor-MATH-1EPOCH

This model is a fine-tuned version of meta-llama/Llama-3.2-3B-Instruct trained on the MATH dataset for one epoch using Intuitor.

Description

Intuitor is a reinforcement learning method introduced in the paper Learning to Reason without External Rewards. It fine-tunes Large Language Models (LLMs) using self-certainty—the model’s own internal confidence—as the sole reward signal.

This approach is part of a novel framework called Reinforcement Learning from Internal Feedback (RLIF), which enables models to learn from intrinsic signals without the need for external rewards, gold labels, or test-case verifiers. Intuitor replaces external rewards in Group Relative Policy Optimization (GRPO) with self-certainty scores, enabling fully unsupervised learning that matches or exceeds standard RL performance on mathematical and coding benchmarks.

Paper: Learning to Reason without External Rewards
Repository: GitHub - sunblaze-ucb/Intuitor

Citation

@article{zhao2025learning,
  title={Learning to Reason without External Rewards},
  author={Zhao, Xuandong and Kang, Zhewei and Feng, Aosong and Levine, Sergey and Song, Dawn},
  journal={arXiv preprint arXiv:2505.19590},
  year={2025}
}