PsyProxy
Reference
Methodology
6 source pages · 1,100 Health Lens dimensions
Native Next routes rebuilt from the verified presentation artifacts: machinery, validation, leaderboard, excitement, topic-model repair notes, transfer, and searchable Health Lens evidence.
Searchable evidence

Health Lens dimension browser

Browse 1,100 Health Lens dimensions with definitions, bands, Excitement Index values, and compact evidence summaries extracted from the verified source pages.

Referencemethodology/psyproxy_machinery.html

PsyProxy Machinery

PsyProxy projects raw text into low-dimensional lenses — each lens is a curated bank of psychometric, social, or behavioral proxies. A regression on these proxies is the PsyProxy "system" you see compete in our benchmarks. This page lists the four lenses we ship today, with dimension counts, training scope, and sample dimensions per lens.

Leaderboardmethodology/gold_medal_page.html

Gold Medal Page

Olympic-style medal tally across the v5 dataset cards. Each (dataset, target) is one competition. Only datasets with at least 5 criterion-model families competing count toward the medal table — uncontested entries (PsyProxy alone) are excluded. PsyProxy competes as one family (best of 4 lenses × strict/permissive + Sinhala translation arms). OpenAI competes as one family (best of Rathje-construct regression on GPT-4o-mini, GPT-4.1-nano, GPT-5-nano). Lexicon-based and topic-model baselines compete as their own singletons. Primary metric per task type: binary =FVE․Binomial, ordinal =Quadratic Kappa, regression =R², multilabel/multiclass =Macro F1.

Methodologymethodology/excitement_index.html

Excitement Index — Methodology

Reference document for the Excitement Index column shown on every dataset card. Authored 2026-04-26 from the 2026-04-25 candidate-metrics literature review and the 2026-04-26 rerun on permissive positives across all four lenses.

Walkthroughwalkthrough/psyproxy_options_and_validation.html

PsyProxy Package Options

Accepts any CSV, re-maps columns to internal names, assigns folds, computes proxy features through a chosen lens, runs the validation pipeline, and writes evaluation artifacts to a user-specified output directory.

Implementation notewalkthrough/topic_model_evaluation_implementation_note.html

What changed in the topic-model pipeline, what failed in the early runs, and why the current scores should be treated as the post-rescue benchmark

This note exists so topic-model evaluation pages can link to one stable explanation instead of repeating fragments of the story. The short version is that we kept the honest early failures, rebuilt the evaluation path so topics are trained and projected split-safely, tuned within train-only pools, removed a sentence-splitting shortcut that made the task artificially easy, and repaired Top2Vec so held-out rows are projected into the learned topic space rather than quietly altering the fitted model.

Transfercontext_validity_transfer_top3.html

PsyProxy Context Validity — Top 3 Cross-Dataset Transfer (strict)

Frame from Larsen et al. (2025) — Context validity = the extent to which a measurement or model trained in one substantive context behaves consistently when applied to a different one.