Victor Veitch
@vveitch.bsky.social
📤 535
📥 72
📝 5
machine learning and artificial intelligence | University of Chicago / Google
come learn about LLM geometry!
add a skeleton here at some point
7 months ago
0
1
0
I'll present this poster tonight at East exhibit hall a-c 2510. 5-7:30 pm. Come chat about alignment!
add a skeleton here at some point
11 months ago
0
7
0
I'll be at NeurIPS Thursday-Sunday; send me an email if you'd like to chat :)
11 months ago
0
2
0
LLM Alignment aims at making model outputs preferred by a ranker while changing as little 'off-target' behavior as possible. Turns out: -best-of-$n$ is the optimal option! -you can contrastively train an LLM to mimic its own best-of-$n$ distribution! BonBon alignment:
arxiv.org/abs/2406.00832
loading . . .
On Spurious Associations and LLM Alignment
Large language models are `aligned' to bias them towards outputting responses that are good on various measures---e.g., we may want them to be helpful, factual, and polite. Often, alignment procedures...
https://simons.berkeley.edu/talks/victor-veitch-university-chicago-2024-11-14
12 months ago
1
6
1
you reached the end!!
feeds!
log in
