README.md

---
license: apache-2.0
library_name: transformers
tags:
- causal-lm
- text-generation
- transformer
- decoder-only
- fixed-embeddings
- binary-token-codes
- research
language:
  - en
---

# Fixed Minimal Binary Code Model

Research checkpoint for the paper:

**Language Models Without a Trainable Input Embedding Table: Learning from Fixed Minimal Binary Token Codes**

## Model variant

This repository contains the **fixed minimal binary token-code model**.

Instead of a trainable input embedding table, each token ID is represented by its exact minimal binary code.

For vocabulary size:

```text
V = 65,536
```

the minimal injective binary code width is:

```text
K = ceil(log2(V)) = 16
```

The 16-dimensional binary code is tiled to model width 1024.

The model therefore uses:

```text
0 trainable input-embedding parameters
```

The output projection remains standard and trainable.

## Architecture

- decoder-only Transformer
- vocabulary size: 65,536
- model width: 1024
- number of layers: 32
- number of attention heads: 32
- context length: 1024
- rotary positional embeddings
- GELU activations
- untied trainable output projection

## Loading example

```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

repo_id = "Bochkov/llm-fix-min-fixed-minimal-binary-code"

tokenizer = AutoTokenizer.from_pretrained(repo_id, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained(repo_id, trust_remote_code=True)
model.eval()

prompt = "Question: What is the capital of France?\nAnswer:"
input_ids = torch.tensor([tokenizer.encode(prompt)], dtype=torch.long)

with torch.no_grad():
    output_ids = model.generate(input_ids, max_new_tokens=3, do_sample=False)

print(tokenizer.decode(output_ids[0].tolist()))
```

## Intended use

This checkpoint is provided for reproducibility of the paper's main claim: a trainable input embedding table is not necessary for useful language modeling in the studied regime.

## Limitations

This model is a research checkpoint. It is not intended for deployment. It may produce incorrect, biased, unsafe, or nonsensical outputs.

## Training data

The model was trained on the same FineWeb-Edu + Cosmopedia mixture used for the matched comparisons in the paper. Dataset terms and licenses are those of the original datasets.

---

## 🧑‍🔬 Citation & Concept

If you use this model or the underlying concepts in your research, please cite our work:

```
@misc{bochkov2026languagemodelstrainableinput,
      title={Language Models Without a Trainable Input Embedding Table: Learning from Fixed Minimal Binary Token Codes}, 
      author={A. Bochkov},
      year={2026},
      eprint={2605.09751},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2605.09751}, 
}
```