<feed xmlns='http://www.w3.org/2005/Atom'>
<title>research.git/q1-prediction-to-reasoning, branch master</title>
<subtitle>jayrup research agenda: how AI models actually work internally</subtitle>
<id>http://git.jayrup.me/c/research.git/atom?h=master</id>
<link rel='self' href='http://git.jayrup.me/c/research.git/atom?h=master'/>
<link rel='alternate' type='text/html' href='http://git.jayrup.me/c/research.git/'/>
<updated>2026-07-23T15:13:39Z</updated>
<entry>
<title>Q1: add 6 attack strategies — distill down, phase transitions, universal geometry, mechanistic embedding, simple data, necessary ablation</title>
<updated>2026-07-23T15:13:39Z</updated>
<author>
<name>CaptainJack2491</name>
<email>jayrupnakawala@gmail.com</email>
</author>
<published>2026-07-23T15:13:39Z</published>
<link rel='alternate' type='text/html' href='http://git.jayrup.me/c/research.git/commit/?id=f161a605dc9addaca7b858d253d3d9da654ed66a'/>
<id>urn:sha1:f161a605dc9addaca7b858d253d3d9da654ed66a</id>
<content type='text'>
</content>
</entry>
<entry>
<title>Add Q1 deep dive: why prediction→reasoning, why scale matters, sub-questions</title>
<updated>2026-07-23T14:51:59Z</updated>
<author>
<name>CaptainJack2491</name>
<email>jayrupnakawala@gmail.com</email>
</author>
<published>2026-07-23T14:51:59Z</published>
<link rel='alternate' type='text/html' href='http://git.jayrup.me/c/research.git/commit/?id=93123f06c07459e6d692d7b37f0f880272a4cd7f'/>
<id>urn:sha1:93123f06c07459e6d692d7b37f0f880272a4cd7f</id>
<content type='text'>
</content>
</entry>
</feed>
