Mercury fullgrid methods paper reveals coverage failure in sparse LLM activation observation, suggesting recovery with dense quantile fullgrid runs.
A Mercury fullgrid methods paper showing that sparse or coarse LLM activation observation can allocate more than 100 million cells while failing a 90% coverage gate, and that dense per-channel quantile fullgrid runs recover auditable coverage across a 19-run archive.
No takes yet. Share an insight, caveat, or question.
Ho Yiing Chen (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: