Instructions to use zeng123/PonderLM-2-Pythia-410m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use zeng123/PonderLM-2-Pythia-410m with Transformers:

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

pipe = pipeline("text-generation", model="zeng123/PonderLM-2-Pythia-410m", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("zeng123/PonderLM-2-Pythia-410m", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("zeng123/PonderLM-2-Pythia-410m", trust_remote_code=True)
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 zeng123/PonderLM-2-Pythia-410m with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "zeng123/PonderLM-2-Pythia-410m"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zeng123/PonderLM-2-Pythia-410m",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/zeng123/PonderLM-2-Pythia-410m

SGLang

How to use zeng123/PonderLM-2-Pythia-410m 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 "zeng123/PonderLM-2-Pythia-410m" \
    --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": "zeng123/PonderLM-2-Pythia-410m",
		"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 "zeng123/PonderLM-2-Pythia-410m" \
        --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": "zeng123/PonderLM-2-Pythia-410m",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use zeng123/PonderLM-2-Pythia-410m with Docker Model Runner:
```
docker model run hf.co/zeng123/PonderLM-2-Pythia-410m
```

PonderLM-2-Pythia-410m

Pythia-410m architecture pretrained with PonderLM-2, the method introduced in PonderLM-2: Pretraining LLM with Latent Thoughts in Continuous Space (ICML 2026 Spotlight).

TL;DR. Chain-of-Thought scales test-time compute by generating extra tokens. PonderLM-2 does the same at pretraining time, but in continuous space: before predicting each next token the model first emits a few latent thoughts — extra last-hidden-state vectors — and feeds them back into itself.

   vanilla:      x₁ ──► x₂ ──► x₃ ──► x₄

   PonderLM-2:   x₁ ──► z₁ ──► x₂ ──► z₂ ──► x₃ ──► z₃ ──► x₄ ──► z₄
                       z_i = latent thought emitted before predicting x_{i+1}

Code: LUMIA-Group/PonderLM-2
Paper: arXiv:2509.23184
Sibling checkpoint: zeng123/PonderLM-2-Pythia-1.4b

Usage

The model ships with a custom modeling_gpt_neox.py that runs the pondering forward pass. Loading via AutoModelForCausalLM requires trust_remote_code=True:

import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

ckpt = "zeng123/PonderLM-2-Pythia-410m"
tok = AutoTokenizer.from_pretrained(ckpt)
model = AutoModelForCausalLM.from_pretrained(
    ckpt,
    torch_dtype=torch.bfloat16,
    trust_remote_code=True,
).cuda()

prompt = "The mitochondria is "
out = model.generate(
    **tok(prompt, return_tensors="pt").to(model.device),
    max_new_tokens=64,
    use_cache=True,
)
print(tok.decode(out[0], skip_special_tokens=True))

Model details


Architecture	GPT-NeoX (Pythia family)
Parameters	410 M
Hidden size	1024
Layers	24
Attention heads	16
Context length	2048
Vocabulary	50 304
Tokenizer	GPT-NeoX BPE (same as Pythia)
Precision	BF16

Citation

@article{zeng2025ponderlm,
  title={Ponderlm-2: Pretraining llm with latent thoughts in continuous space},
  author={Zeng, Boyi and Li, He and Song, Shixiang and Wang, Yixuan and Wang, Zitong and He, Ziwei and Wang, Xinbing and Lin, Zhouhan},
  journal={arXiv preprint arXiv:2509.23184},
  year={2025}
}

Acknowledgements

Built on top of the Pythia training stack and LLaMA-Factory. The PonderLM baseline implementation is adapted from LUMIA-Group/PonderingLM.

Downloads last month: 20

Safetensors

Model size

0.4B params

Tensor type

BF16

Dataset used to train zeng123/PonderLM-2-Pythia-410m

Paper for zeng123/PonderLM-2-Pythia-410m

PonderLM-2: Pretraining LLM with Latent Thoughts in Continuous Space

Paper • 2509.23184 • Published Sep 27, 2025 • 2