README.md

---
license: mit
datasets:
- dleemiller/wiki-sim
- sentence-transformers/stsb
language:
- en
metrics:
- spearmanr
- pearsonr
base_model:
- answerdotai/ModernBERT-base
pipeline_tag: text-classification
library_name: sentence-transformers
tags:
- cross-encoder
- modernbert
- sts
- stsb
- stsbenchmark-sts
model-index:
- name: CrossEncoder based on answerdotai/ModernBERT-base
  results:
  - task:
      type: semantic-similarity
      name: Semantic Similarity
    dataset:
      name: sts test
      type: sts-test
    metrics:
    - type: pearson_cosine
      value: 0.9162245947821821
      name: Pearson Cosine
    - type: spearman_cosine
      value: 0.9121555789491528
      name: Spearman Cosine
  - task:
      type: semantic-similarity
      name: Semantic Similarity
    dataset:
      name: sts dev
      type: sts-dev
    metrics:
    - type: pearson_cosine
      value: 0.9260833551026787
      name: Pearson Cosine
    - type: spearman_cosine
      value: 0.9236030687487745
      name: Spearman Cosine
---
# ModernBERT Cross-Encoder: Semantic Similarity (STS)

Cross encoders are high performing encoder models that compare two texts and output a 0-1 score.
I've found the `cross-encoders/roberta-large-stsb` model to be very useful in creating evaluators for LLM outputs.
They're simple to use, fast and very accurate.

Like many people, I was excited about the architecture and training uplift from the ModernBERT architecture (`answerdotai/ModernBERT-base`).
So I've applied it to the stsb cross encoder, which is a very handy model. Additionally, I've added
pretraining from a much larger semi-synthetic dataset `dleemiller/wiki-sim` that targets this kind of objective.
The inference performance efficiency, expanded context and simplicity make this a really nice platform as an evaluator model.

---

## Features
- **High performing:** Achieves **Pearson: 0.9162** and **Spearman: 0.9122** on the STS-Benchmark test set.
- **Efficient architecture:** Based on the ModernBERT-base design (149M parameters), offering faster inference speeds.
- **Extended context length:** Processes sequences up to 8192 tokens, great for LLM output evals.
- **Diversified training:** Pretrained on `dleemiller/wiki-sim` and fine-tuned on `sentence-transformers/stsb`.

---

## Performance

| Model                          | STS-B Test Pearson | STS-B Test Spearman | Context Length | Parameters | Speed  |
|--------------------------------|--------------------|---------------------|----------------|------------|---------|
| `ModernCE-large-sts`           | **0.9256**         | **0.9215**          | **8192**       | 395M       | **Medium** |
| `ModernCE-base-sts`            | **0.9162**         | **0.9122**          | **8192**       | 149M       | **Fast** |
| `stsb-roberta-large`           | 0.9147            | -              | 512            | 355M       | Slow    |
| `stsb-distilroberta-base`      | 0.8792            | -              | 512            | 82M        | Fast    |


---

## Usage

To use ModernCE for semantic similarity tasks, you can load the model with the Hugging Face `sentence-transformers` library:

```python
from sentence_transformers import CrossEncoder

# Load ModernCE model
model = CrossEncoder("dleemiller/ModernCE-base-sts")

# Predict similarity scores for sentence pairs
sentence_pairs = [
    ("It's a wonderful day outside.", "It's so sunny today!"),
    ("It's a wonderful day outside.", "He drove to work earlier."),
]
scores = model.predict(sentence_pairs)

print(scores)  # Outputs: array([0.9184, 0.0123], dtype=float32)
```

### Output
The model returns similarity scores in the range `[0, 1]`, where higher scores indicate stronger semantic similarity.

---

## Training Details

### Pretraining
The model was pretrained on the `pair-score-sampled` subset of the [`dleemiller/wiki-sim`](https://huggingface.co/datasets/dleemiller/wiki-sim) dataset. This dataset provides diverse sentence pairs with semantic similarity scores, helping the model build a robust understanding of relationships between sentences.
- **Classifier Dropout:** a somewhat large classifier dropout of 0.3, to reduce overreliance on teacher scores.
- **Objective:** STS-B scores from `cross-encoder/stsb-roberta-large`.

### Fine-Tuning
Fine-tuning was performed on the [`sentence-transformers/stsb`](https://huggingface.co/datasets/sentence-transformers/stsb) dataset.

### Validation Results
The model achieved the following test set performance after fine-tuning:
- **Pearson Correlation:** 0.9162
- **Spearman Correlation:** 0.9122

---

## Model Card

- **Architecture:** ModernBERT-base
- **Tokenizer:** Custom tokenizer trained with modern techniques for long-context handling.
- **Pretraining Data:** `dleemiller/wiki-sim (pair-score-sampled)`
- **Fine-Tuning Data:** `sentence-transformers/stsb`

---

## Thank You

Thanks to the AnswerAI team for providing the ModernBERT models, and the Sentence Transformers team for their leadership in transformer encoder models.

---

## Citation

If you use this model in your research, please cite:

```bibtex
@misc{moderncestsb2025,
  author = {Miller, D. Lee},
  title = {ModernCE STS: An STS cross encoder model},
  year = {2025},
  publisher = {Hugging Face Hub},
  url = {https://huggingface.co/dleemiller/ModernCE-base-sts},
}
```

---

## License

This model is licensed under the [MIT License](LICENSE).