smds 1.0.0

A plug-and-play, scikit-learn compatible implementation of Supervised Multi-Dimensional Scaling (SMDS) for automatic feature manifold discovery in LLMs.

Supervised Multi-Dimensional Scaling

This is a stand-alone implementation of Supervised Multi-Dimensional Scaling (SMDS) from the paper "Shape Happens: Automatic Feature Manifold Discovery in LLMs". It contains a plug-and-play class written with the familiar scikit-learn interface. SMDS supports several template shapes to discover manifolds of various forms.

Contact person: Federico Tiblias

UKP Lab | TU Darmstadt

Don't hesitate to report an issue if you have further questions or spot a bug.

Getting started

With uv (recommended):

uv add supervised-multidimensional-scaling

With pip:

pip install supervised-multidimensional-scaling

Usage

The SupervisedMDS class provides a scikit-learn style interface that is straightforward to use. Unlike standard MDS, it requires a target manifold shape (e.g., ClusterShape, CircularShape) to define the ideal geometry.

Fit & Transform

You can instantiate the model, fit it to data (X, y), and transform your input into a low-dimensional embedding:

import numpy as np
from smds import SupervisedMDS
from smds.shapes import ClusterShape

# Example data
X = np.random.randn(100, 20)   # 100 samples, 20 features
y = np.random.randint(0, 5, size=100)  # Discrete labels (clusters)

# Instantiate and fit
# manifold can be any class inheriting from BaseShape
smds = SupervisedMDS(n_components=2, manifold=ClusterShape(), alpha=0.1)
smds.fit(X, y)

# Transform to low-dimensional space
X_proj = smds.transform(X)
print(X_proj.shape)  # (100, 2)

Manifold Discovery

The library provided a high level pipeline to automatically discover the intrinsic manifold of your data. The discover_manifolds utility evaluates a set of hypothesis shapes (for example: clusters, circles, hierarchies) and ranks them based on how well they explain the data structure using cross validation.

from smds.pipeline.discovery_pipeline import discover_manifolds
from smds.pipeline import open_dashboard

# Run discovery pipeline
# Evaluates default shapes (Cluster, Circular, Hierarchical, etc.)
# Returns a DataFrame sorted by best fit (lowest stress / highest score)
df_results, save_path = discover_manifolds(
    X, 
    y, 
    smds_components=2,           # Target dimensionality
    n_folds=5,                   # Cross-validation folds
    experiment_name="My_Exp",    # Name for saved results
    n_jobs=-1                    # Use all available cores
)

print(f"Best matching shape: {df_results.iloc[0]['shape']}")
print(df_results.head())

# Launch the interactive Streamlit dashboard to explore results and plots
open_dashboard.main(save_path)

The discovery pipeline handles:

• Hypothesis Testing: Iterates through a default or custom list of manifold shapes.
• Cross-Validation: Uses k-fold CV to ensure robust scoring.
• Caching: Caches intermediate results to resume interrupted experiments.
• Visualization: Generates interactive plots for the dashboard.

Development

This seciton is especially usefull if you consider contributing to the library!

Documentation

To build and serve the documentation locally:

mkdocs serve

[!NOTE] The dev dependency group includes heavy libraries such as torch and transformers.

Testing

Run the test suite using pytest:

make test

Cite

Please use the following citation:

@misc{tiblias2025shapehappensautomaticfeature,
      title={Shape Happens: Automatic Feature Manifold Discovery in LLMs via Supervised Multi-Dimensional Scaling}, 
      author={Federico Tiblias and Irina Bigoulaeva and Jingcheng Niu and Simone Balloccu and Iryna Gurevych},
      year={2025},
      eprint={2510.01025},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2510.01025}, 
}