Integrated Gradients

Bases: MultitaskExplainerMixin, AttributionExplainer

Integrated Gradients (IG) is a gradient-based interpretability method that attributes importance scores to input features (e.g., tokens) by integrating the model’s gradients along a path from a baseline input to the actual input.

The method is designed to address some of the limitations of standard gradients, such as saturation and noise, by averaging gradients over interpolated inputs rather than relying on a single local gradient.

Reference: Sundararajan et al. (2017). Axiomatic Attribution for Deep Networks. Paper

Examples:

>>> from interpreto import IntegratedGradients
>>> method = IntegratedGradients(model=model, tokenizer=tokenizer,
>>>                              batch_size=4, n_interpolations=50)
>>> explanations = method.explain(model_inputs=text)

Parameters:

Name	Type	Description	Default
model	PreTrainedModel	model to explain	required
tokenizer	PreTrainedTokenizer	Hugging Face tokenizer associated with the model	required
batch_size	int	batch size for the attribution method	4
device	device	device on which the attribution method will be run	None
inference_mode	Callable[[Tensor], Tensor]	The mode used for inference. It can be either one of LOGITS, SOFTMAX, or LOG_SOFTMAX. Use InferenceModes to choose the appropriate mode.	LOGITS
n_interpolations	int	the number of interpolations to generate	10
baseline	Tensor | float | None	the baseline to use for the interpolations	None