---
base_model: meta-llama/Llama-3.1-8B-Instruct
license: llama3
pipeline_tag: text-generation
tags:
- biology
- protein
- molecule
- dna
- rna
- multimodal
- structure-grounded
---

# Cuttlefish

**Cuttlefish** is a unified all-atom multimodal LLM that grounds language reasoning in geometric cues while scaling structural tokens with structural complexity. Built on Llama-3.1-8B-Instruct, it extends the base LLM with a graph encoder and a Scaling-Aware Patching connector for processing proteins, molecules, DNA, and RNA structures.

The model was introduced in the paper [Scaling-Aware Adapter for Structure-Grounded LLM Reasoning](https://huggingface.co/papers/2602.02780).

**Code**: [https://github.com/zihao-jing/Cuttlefish](https://github.com/zihao-jing/Cuttlefish)

## Quick start

To use the model, you can download the weights using `huggingface_hub`. Running inference requires the original codebase.

```python
from huggingface_hub import snapshot_download

# Download model
local_dir = snapshot_download("zihaojing/Cuttlefish")

# Run inference (requires cuttlefish codebase)
# python src/runner/inference.py --config configs/inference/octopus_8B_s3_v1_5.yaml
```

## Input format

Cuttlefish accepts a unified parquet schema with structural graph columns:

| Field | Description |
|---|---|
| `modality` | `"molecule"`, `"protein"`, `"dna"`, or `"rna"` |
| `node_feat` | Atom/node features (N × d) |
| `pos` | 3D coordinates in Å (N × 3) |
| `edge_index` | Spatial graph edges in COO (2 × E) |
| `messages` | Chat-style instruction with `<STRUCTURE>` token |

The `<STRUCTURE>` placeholder in the user message is replaced by the encoded structural tokens at inference time.

## Training details

- **Base model**: Llama-3.1-8B-Instruct
- **Encoder**: [Cuttlefish-Encoder](https://huggingface.co/zihaojing/Cuttlefish-Encoder) (pretrained on all-atom graph data)
- **SFT data**: [Cuttlefish-SFT-Data](https://huggingface.co/datasets/zihaojing/Cuttlefish-SFT-Data)
- **Training stages**: 2-stage SFT — connector training then full LLM fine-tuning with LoRA

## Related resources

| Resource | Link |
|---|---|
| Cuttlefish-Encoder | [zihaojing/Cuttlefish-Encoder](https://huggingface.co/zihaojing/Cuttlefish-Encoder) |
| SFT instruction data | [zihaojing/Cuttlefish-SFT-Data](https://huggingface.co/datasets/zihaojing/Cuttlefish-SFT-Data) |
| Encoder pretraining data | [zihaojing/Cuttlefish-Encoder-Data](https://huggingface.co/datasets/zihaojing/Cuttlefish-Encoder-Data) |

## Citation

```bibtex
@article{jing2026cuttlefish,
  title     = {Cuttlefish: Scaling-Aware Adapter for Structure-Grounded LLM Reasoning},
  author    = {Jing, Zihao and Zeng, Qiuhao and Fang, Ruiyi and Li, Yan Yi and Sun, Yan and Wang, Boyu and Hu, Pingzhao},
  booktitle = {Proceedings of the 43rd International Conference on Machine Learning (ICML)},
  year      = {2026},
  url       = {https://arxiv.org/abs/2602.02780}
}
```