PhilBench

An interactive dashboard comparing the philosophical views of large language models to the consensus of professional philosophers, using the 100 questions of the 2020 PhilPapers Survey (David Bourget & David J. Chalmers; 1,785 respondents).

Each model answers all 100 questions across multiple prompt variants and runs; the dashboard lets you browse per-model and per-question results, track trends over release date and capability, and compare model answers to the philosopher baseline.

Hosting

This repository hosts the static dashboard via GitHub Pages. index.html lazily fetches its data from data/*.json, so it must be served over HTTP (it will not work opened directly from file://). GitHub Pages serves it over HTTPS, which satisfies that requirement.

index.html              # dashboard shell
data/core.json          # per-model / per-question results (loaded on start)
data/consistency.json   # View Consistency data (loaded on demand)
data/variants.json      # Prompt Variants data (loaded on demand)

Eval code and the full response data bundles live separately.