← All parts of this equation

Equation 34 · Part 3 · A Chatbot Confessed to Being Built by a Company That Never Trained It

Symbol p_pre

s=ln⁡(ppost/(1−ppost)ppre/(1−ppre))s=\ln\left(\frac{p_{\mathrm{post}}/(1-p_{\mathrm{post}})}{p_{\mathrm{pre}}/(1-p_{\mathrm{pre}})}\right)
pprep_{\mathrm{pre}}

What this part means

ppp_pre occurs below the fraction bar. The quantity above the bar is divided by this expression; zero is excluded as a denominator.

Its job in the formula

ppp_pre occurs below the fraction bar. The quantity above the bar is divided by this expression; zero is excluded as a denominator.

The passage around this formula

…measures how a curation step — a filter, a quality score, a deduplication pass — changes the trait’s prevalence between the raw corpus the source produced, pprep_{\mathrm{pre}} , and the corpus that actually reaches a downstream trainer, ppostp_{\mathrm{post}} : s>0 means curation enriches for the trait, s<0 means curation purges it, and s=0 means curation is neutral toward it. None of these three quantities requires the exposed and unexposed models to share a…

Read this part in the article →

Learn the underlying idea

A subscript is a label attached below a symbol. It often selects a time step, component, category, or member of a sequence.

Open the illustrated subscripts: which member of a family? guide →

See this notation across published equations →

The article lists its research sources here.