Site profile
ablation-bench.github.io
Built with prismjs 1.30.0, docsify 4, docsify-themeable 0, docsify-tabs 1, docsify-copy-code 2.1.1. 1 page and 10 stored files, last crawled 2026-10-03.
5 libraries 0 identifiers 3 third-party hosts 1 crawl open the site
Technologies and libraries
Recognised by fingerprinting the stored files and bundles against known releases; a version is the release the bytes match.
| Library | Version | Seen |
|---|---|---|
| prismjs | 1.30.0 | 2026-10-03 |
| docsify | 4 | 2026-10-03 |
| docsify-themeable | 0 | 2026-10-03 |
| docsify-tabs | 1 | 2026-10-03 |
| docsify-copy-code | 2.1.1 | 2026-10-03 |
Tracking, tag and verification IDs
No analytics property, tag manager container, pixel or site-verification token was found in the stored source.
Third-party hosts
Hosts outside ablation-bench.github.io that its pages load scripts, frames or stylesheets from, and that its own scripts name in absolute URLs (connect). Each links to every site loading from the same host.
| Host | Loaded as | Found in |
|---|---|---|
| cdn.jsdelivr.net | script stylesheet | view source |
| gradio.s3-us-west-2.amazonaws.com | script | view source |
| unpkg.com | script | view source |
Stored pages and scripts
The newest stored version of each file, newest first@if (p.FilesTruncated) {
Timeline
One entry per crawl, newest first, with what changed since the crawl before: libraries, identifiers, third-party hosts and files. Historical versions stay viewable because the stored chunks are shared between versions, never copied.
-
first crawl
1 page, 10 files, 2.4 KB; 6 libraries, 0 identifiers, 3 third-party hosts
6 libraries added
- docsify 4
- docsify-themeable 0
- docsify-tabs 1
- docsify-copy-code 2
- prismjs 1.30.0
- docsify-copy-code 2.1.1
4 third-party hosts added
10 files added
- https://ablation-bench.github.io/ html
- https://cdn.jsdelivr.net/npm/docsify@4/lib/docsify.min.js js
- https://cdn.jsdelivr.net/npm/docsify-themeable@0…/js/docsify-themeable.min.js js
- https://cdn.jsdelivr.net/npm/docsify-tabs@1 js
- https://unpkg.com/docsify-copy-code@2 js
- https://cdn.jsdelivr.net/npm/docsify@4/lib/plugins/zoom-image.min.js js
- https://cdn.jsdelivr.net/npm/docsify@4/lib/plugins/search.js js
- https://cdn.jsdelivr.net/npm/[email protected]/components/prism-json.min.js js
- https://cdn.jsdelivr.net/npm/[email protected]/components/prism-latex.min.js js
- https://unpkg.com/[email protected]/dist/docsify-copy-code.min.js js
Questions about ablation-bench.github.io
- How does PageSourceSearch know what ablation-bench.github.io is built with?
- From the site's own source code. The crawler stores the exact bytes of its pages and first-party scripts; the libraries are recognised by fingerprinting that code against known releases, the identifiers are read out of the tag snippets, and the third-party hosts are the script, iframe and stylesheet sources in the HTML and the absolute URLs inside the scripts. Nothing is inferred from headers or guessed.
- How far back does the history of ablation-bench.github.io go?
- To the first crawl the timeline lists. Every crawl records what the site looked like; when a file's content changes, its earlier version stays viewable because the stored chunks are never deleted, only mapped. A crawl that finds a file unchanged adds no copy.
- Can I see an older version of a script from ablation-bench.github.io?
- Yes. In the stored files list, every date under a file opens the version that was live at that crawl, with the same viewer as the current one. The raw bytes can be downloaded from there.
- Which other websites use the same tracking IDs as ablation-bench.github.io?
- Each identifier links to its reverse lookup, the list of every site in the index whose source carries the same id: the sites one analytics property or tag manager container is shared across.
Look up another site, browse sites by tracking ID, third-party host or technology. Site owners: see the crawler page for how stored pages are removed.