Home/ Tool index/ Network-Grounded Truth Sieve
Launch ready Cross-session Interactive · live Wikipedia

Network-Grounded Truth Sieve

A model predicts fluent words, not true ones. This sieve strips a passage to its proper-noun phrases and dates and runs a live Wikipedia audit, asking one thing of each anchor: does it actually exist?

tools/network-grounded-truth-sieve/

Live preview · launch for the interactive version

§ A · What it makes visible

three hidden mechanisms
Fig. 01

The token sieve

Highlight the hard factual anchors and dim the fluency that hides them.

Fig. 02

Live existence check

Each anchor is queried against Wikipedia — invented names fail here.

Fig. 03

Existence ≠ truth

A passage can pass existence and still be false — that's the next sieve's job.

§ B · How to investigate it

run it like an experiment, not a toy
01 · Predict

Before the audit

Predict which anchors in the passage are real.

guess: the general is invented
02 · Change one thing

Swap one entity

Replace a real name with an invented one; re-run.

'Versailles' → 'Shady Pines'
03 · Compare evidence

Read the log

Which anchors resolved, which came back unfound?

fabricated anchor fails
04 · Name it

Name the gap

Say what failed: a fabricated entity, or a real but unverifiable one.

fabrication = no record

§ C · Debrief questions

after the investigation
Why check whole phrases instead of single words?
Can a passage be all-real and still be a lie?
What does fluency hide from a casual reader?
When is 'it exists' not enough?

§ D · Related

pairs well with · use in context