From Disclosure to Substance: The Next Step for AI Transparency in Science

Disclosure + opinion figures — reproduction bundle================================================== Python scripts-------------- generateₒpinionfigures. py — Figure 1 (temporal + forest) & Figure 2 (field-specific ORs), PDFs granularityₐnalysis. py — tier analysis: forest PNG, tier distribution PNG, modeling frame CSV granularitybootstrap. py — bootstrap Data/Code OR violin PNGs + JSON Data under output/ (bundled) ---------------------------- scorecard₂024₅000. csv, scorecard₂025₅000. csv papers₂024₅000. csv, papers₂025₅000. csv granularityₘodelingframe. csv (for bootstrap without re-fetching full text) opinionfigures/fieldᵣesults. csv (optional snapshot from a prior run of the field models) Setup----- pip install -r requirements. txt Opinion figures (PDF, as in the manuscript) ------------------------------------------- cd /path/to/disclosuregranularityfigures python3 generateₒpinionfigures. py Writes: output/opinionfigures/figure1combined. pdf output/opinionfigures/figure2fieldₛpecific. pdf output/opinionfigures/fieldᵣesults. csv (refit from scorecards + papers) Note: Figure 1 panels (A) bar heights and (B) pooled ORs are the values coded in generateₒpinionfigures. py (main-text point estimates). Figure 2 is estimated from the bundled scorecard + papers tables (fields with n ≥ 250). Granularity + bootstrap figures (PNG) ------------------------------------- python3 granularitybootstrap. py # fast if granularityₘodelingframe. csv present python3 granularityₐnalysis. py # optional: refetch/cache full text, then PNGs Typical outputs: output/granularityforestₜiers. png output/granularityₜierdistribution. png output/granularitybootstrapdataₒr. png output/granularitybootstrapcodeₒr. png Reference PNGs from an earlier run: output/figuresᵣeference/ Environment (optional) ---------------------- GRANULARITYWORKERS=8 GRANULARITYMAXN=0 GRANULARITYBOOTSTRAPB=2000 GRANULARITYBOOTSTRAPSTRATIFIED=1