Research Radar · Experimental implementation
Confabulation Scaling
Scaling-law experiments for factual-recall errors in language models.
What it investigates
Confabulation Scaling studies a workflow for relating topic frequency and model scale to factual-recall errors.
Implementation record
The implementation documents a corpus-frequency estimator, reference-data downloader, index builder, citation checker, and sigmoid fitter. Together they support estimating topic prevalence, checking cited material, and fitting a curve for exploratory recall predictions.
Public record
The linked research item is the public source recorded for this Research Radar build. It provides context for the implementation; this page does not claim to reproduce its results. The repository and its documentation are the current public materials for inspecting the build itself.
Scope and limitations
This is an experimental analysis toolkit. It does not establish a universal law for model errors, validate every corpus or citation, or make a predictive guarantee for a particular language model.