Thomas Spooks 🎃👻💀

thomasfuchs.at

19 January 2025 04:12

Still astounded how people think that the government force-closing a social network that's literally used by 170 million people in the US is just fine or even good. It would not surprise me if they go after VPNs next. They tried this already: en.wikipedia.org/wiki/RESTRIC...

Laura

lauraruis.bsky.social

21 November 2024 21:24

To my surprise, we find the opposite of what I thought when we started this project: The approach to reasoning LLMs use looks unlike retrieval, and more like a generalisable strategy synthesising procedural knowledge from many documents doing a similar form of reasoning.

Laura

lauraruis.bsky.social

21 November 2024 21:23

How do LLMs learn to reason from data? Are they ~retrieving the answers from parametric knowledge🦜? In our new preprint, we look at the pretraining data and find evidence against this: Procedural knowledge in pretraining drives LLM reasoning ⚙️🔢 🧵⬇️

DotKrueger

dotkrueger.bsky.social

21 November 2024 21:17

The absence of verified profiles makes this almost unusable

Kapalı içerik