Skip to content

[P1] Speed up training of multiple DAS interventions with caching #76

@aryamanarora

Description

@aryamanarora

Commonly, we want to exhaustively train DAS on every layer and position (or e.g. every attention head in a layer) to find which ones are causally relevant for the model's computations. When dealing with a fixed dataset, we could speed this process up by caching and reusing activations. Unclear what the best way to implement this is; should already be possible to have a minimal example with CollectIntervention and the activations_sources= arg in inference.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions