Saved in:
Bibliographic Details
Main Authors: Saeed, Muhammed, Razniewski, Simon
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.24080
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914420567834624
author Saeed, Muhammed
Razniewski, Simon
author_facet Saeed, Muhammed
Razniewski, Simon
contents Benchmarks such as MMLU suggest flagship language models approach factuality saturation, with scores above 90\%. We show this picture is incomplete. \emph{LLMpedia} generates encyclopedic articles entirely from parametric memory, producing ${\sim}$1M articles across three model families without retrieval. For gpt-5-mini, the verifiable true rate on Wikipedia-covered subjects is only 74.7\% -- more than 15 percentage points below the benchmark-based picture, consistent with the availability bias of fixed-question evaluation. Beyond Wikipedia, frontier subjects verifiable only through curated web evidence fall further to 63.2\% true rate. Wikipedia covers just 61\% of surfaced subjects, and three model families overlap by only 7.3\% in subject choice. In a capture-trap benchmark inspired by prior analysis of Grokipedia, LLMpedia achieves substantially higher factuality at roughly half the textual similarity to Wikipedia. Unlike Grokipedia, every prompt, artifact, and evaluation verdict is publicly released, making LLMpedia the first fully open parametric encyclopedia -- bridging factuality evaluation and knowledge materialization. All data, code, and a browsable interface are at https://llmpedia.net.
format Preprint
id arxiv_https___arxiv_org_abs_2603_24080
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle LLMpedia: A Transparent Framework to Materialize an LLM's Encyclopedic Knowledge at Scale
Saeed, Muhammed
Razniewski, Simon
Computation and Language
Databases
Benchmarks such as MMLU suggest flagship language models approach factuality saturation, with scores above 90\%. We show this picture is incomplete. \emph{LLMpedia} generates encyclopedic articles entirely from parametric memory, producing ${\sim}$1M articles across three model families without retrieval. For gpt-5-mini, the verifiable true rate on Wikipedia-covered subjects is only 74.7\% -- more than 15 percentage points below the benchmark-based picture, consistent with the availability bias of fixed-question evaluation. Beyond Wikipedia, frontier subjects verifiable only through curated web evidence fall further to 63.2\% true rate. Wikipedia covers just 61\% of surfaced subjects, and three model families overlap by only 7.3\% in subject choice. In a capture-trap benchmark inspired by prior analysis of Grokipedia, LLMpedia achieves substantially higher factuality at roughly half the textual similarity to Wikipedia. Unlike Grokipedia, every prompt, artifact, and evaluation verdict is publicly released, making LLMpedia the first fully open parametric encyclopedia -- bridging factuality evaluation and knowledge materialization. All data, code, and a browsable interface are at https://llmpedia.net.
title LLMpedia: A Transparent Framework to Materialize an LLM's Encyclopedic Knowledge at Scale
topic Computation and Language
Databases
url https://arxiv.org/abs/2603.24080