langspace.probe.arithmetic package

Module contents

class langspace.probe.arithmetic.ArithmeticProbe(model: LangVAE, data: List[Tuple[Sentence, Sentence]], ops: List[ArithmeticOps], annotations: Dict[str, List[str]] = None)

Bases: LatentSpaceProbe

A probe for exploring arithmetic operations in the latent space of a language model variational autoencoder (LM-VAE).

This class applies specified arithmetic operations to latent representations obtained from pairs of sentences. It supports operations such as summation, subtraction, and averaging. In addition, the probe can generate a report in the form of a pandas DataFrame summarizing the original source and target sentences alongside the results of the applied operations.

model

The LM-VAE model to be probed.

Type:

LangVAE

data

A list of sentence pairs as (source, target) tuples.

Type:

List[Tuple[Sentence, Sentence]]

ops

A list of arithmetic operations to be applied to the latent vectors.

Type:

List[ArithmeticOps]

annotations

Dictionary of annotation types to be processed and all their

Type:

Dict[str, List[str]], optional

possible values.

arithmetic(source: Tensor, target: Tensor) → List[Tensor]

Apply arithmetic operations to the source and target latent representations.

Parameters:

• source (Tensor) – The latent representation of the source sentences.

• target (Tensor) – The latent representation of the target sentences.

Returns:

A list of tensors, each resulting from applying the corresponding arithmetic operation.

report() → DataFrame

Generate a report summarizing the arithmetic probe results.

The final DataFrame will have the following columns:

• source: The original source sentence surfaces.

• target: The original target sentence surfaces.

• op: The arithmetic operation applied (as a lowercase string).

• generate: The generated sentence after applying the latent operation.

Returns:

A DataFrame containing a detailed report for each arithmetic operation.