| --- |
| license: other |
| library_name: transformers |
| pipeline_tag: text-generation |
| --- |
| |
| # Steerling-8B |
|
|
| Steerling-8B is an interpretable causal diffusion language model introduced in the paper [Scaling Inherently Interpretable Language Models](https://huggingface.co/papers/2608.07594). |
|
|
| - Project page: [https://www.guidelabs.ai/papers/scaling-inherently-interpretable-language-models/](https://www.guidelabs.ai/papers/scaling-inherently-interpretable-language-models/) |
| - Code: [https://github.com/guidelabs/steerling](https://github.com/guidelabs/steerling) |
|
|
| ## Usage |
|
|
| ```python |
| import torch |
| from transformers import AutoModel, AutoTokenizer |
| from steerling import SteerlingGenerator, GenerationConfig |
| |
| model_id = "guidelabs/steerling-8b" |
| model = AutoModel.from_pretrained(model_id, trust_remote_code=True, dtype=torch.bfloat16) |
| tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True) |
| generator = SteerlingGenerator.from_model(model, tokenizer, device="cuda") |
| |
| prompt = "The key to understanding neural networks is" |
| config = GenerationConfig(max_new_tokens=128, steps=128) |
| text = generator.generate(prompt, config) |
| print(text) |
| ``` |
