at://
/
app.bsky.feed.post
/
3moe57g6ahs2v
sign in
All
4
Record
2
Post
1
PostEmbed
1
Post
by @danabra.mov
PostEmbed
by @danabra.mov
Record
by @jimpick.com
Record
by @atsui.org
+ new component
Post
changing an LLM's behavior without retraining the whole model can seem a bit far-fetched. pere martra argues otherwise. and it's all done through neuron pruning. the breakdown: hubs.la/Q04ltRl-0 his book, rearchitecting LLMs: hubs.la/Q04ltRvl0
3d
manning publications
