When Grok answers a question using web search, I sometimes see a source that looks relevant but isn’t very clear about which part of the page supports the response. Does Grok usually read the full page, or might it be working from search result snippets when a site blocks access or loads content dynamically?

I’m trying to judge how much confidence to put in answers based on a source link alone. If you’ve compared Grok’s response with the actual page, have you found a way to tell when it used the page itself rather than a short preview?