WorkspaceBench: Evaluating Interpretability Methods for the Global Workspace

T2Covered by 1 source · first reported by AI Alignment Forum / LessWrong (curated) at 24 Sept, 02:19 UTC

TL;DR We introduce WorkspaceBench, a set of evaluations for how well an activation-to-text tool can read the contents of the “global workspace” of a model, i.e. the intermediate variables during a forward pass. The benchmark comprises 3,356 questions across 27 eval families, spanning topics in…

ai

Sources

• T2WorkspaceBench: Evaluating Interpretability Methods for the Global WorkspaceAI Alignment Forum / LessWrong (curated)24 Sept, 02:19