Rean Fernandes
@reanfds.bsky.social
Large Language models, control systems, mushrooms and life in general :)
@ALU-Freiburg, DE
@ALU-Freiburg, DE
Reading The Hitchhiker’s Guide to Testing Statistical Significance in NLP (aclanthology.org/P18-1128.pdf) and it confirms what I’ve felt: so much of LLM evaluation, especially accuracy metrics, feels like a vibe-check. Wish I’d taken high school stats more seriously :0
a man 's head is surrounded by a circle of light and stars .
ALT: a man 's head is surrounded by a circle of light and stars .
media.tenor.com
December 4, 2024 at 1:56 PM
Reading The Hitchhiker’s Guide to Testing Statistical Significance in NLP (aclanthology.org/P18-1128.pdf) and it confirms what I’ve felt: so much of LLM evaluation, especially accuracy metrics, feels like a vibe-check. Wish I’d taken high school stats more seriously :0
Reposted by Rean Fernandes
Lots of "o1-washing" going on in open model releases these days. CoT, reflection tokens, a sprinkle of RL, all are not o1.
November 18, 2024 at 3:40 PM
Lots of "o1-washing" going on in open model releases these days. CoT, reflection tokens, a sprinkle of RL, all are not o1.
Reposted by Rean Fernandes
Releasing two trillion tokens in the open. huggingface.co/blog/Pclangl...
November 13, 2024 at 5:59 PM
Releasing two trillion tokens in the open. huggingface.co/blog/Pclangl...
When I’m not pulling my hair out trying to read NLP papers or working like Sisyphus on an incredibly obfuscated,codebase that I wrote down in a coffee addled bender months ago, I like to make terrariums. This baby was born 8 days ago, and she’s so beautiful. My friend Jake named her Penelope 🤗
November 10, 2024 at 5:24 PM
When I’m not pulling my hair out trying to read NLP papers or working like Sisyphus on an incredibly obfuscated,codebase that I wrote down in a coffee addled bender months ago, I like to make terrariums. This baby was born 8 days ago, and she’s so beautiful. My friend Jake named her Penelope 🤗