If you train an LLM on only formally verified code, it should not be expected to generate formally verified code.Similarly, if you train an LLM on only published ScholarlyArticles [&#x27;s abstracts], it should not be expected to generate publishable or true text.Traceability for Retraction would be necessary to prevent lossy feedback.

do you plan on adding descriptions of figures and tables?

For now , yes - abstracts and other metadata

you&#x27;re right - I hadn&#x27;t noticed! I fixed it now, thanks for pointing it out

The example you give doesn&#x27;t seem to work - the raw txt does not have authors.

whoa - i haven&#x27;t yet played with MCP - might be a good first project!

This would be awesome wrapped in an MCP server&#x2F;tool call :)

Was super excited that it was going to be the actual papers, kinda cool but just being abstracts doesn&#x27;t go very far, good luck getting the papers working thats gonna be pretty cool once working, then to feed it all into a vector db XD

Just change arxiv.org to arxiv-txt.org in the URL to get the paper info in markdownExample:Original URL:
<a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1706.03762" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1706.03762</a>Change to:
<a href="https:&#x2F;&#x2F;arxiv-txt.org&#x2F;abs&#x2F;1706.03762" rel="nofollow">https:&#x2F;&#x2F;arxiv-txt.org&#x2F;abs&#x2F;1706.03762</a>To fetch the raw text directly, use <a href="https:&#x2F;&#x2F;arxiv-txt.org&#x2F;raw&#x2F;abs&#x2F;1706.03762" rel="nofollow">https:&#x2F;&#x2F;arxiv-txt.org&#x2F;raw&#x2F;abs&#x2F;1706.03762</a>, this will be particularly useful for APIs and agents

Show HN: ArXiv-txt, LLM-friendly ArXiv papers