Work

Coverage Audits for LLM Interpretability Claims

Year: 2026

Type: preprint

Abstract: A Mercury methods note arguing that sparse probing, activation patching, subspace intervention, and pruning claims should report a coverage audit before making broad localization claims. The audit uses 19 fullgrid manifests and shows a 1/7 pass rate for older sparse or coarse protocols versus 12/12 for per-channel quantile dense runs.

Source: Zenodo (CERN European Organization for Nuclear Research)

DOI: 10.5281/zenodo.20729123

Author: Chen, Ho Yiing (norika)

Language: English

Topic: Explainable Artificial Intelligence (XAI)

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W7165106975

Cites:

Cited by:

Related to: