Very well deserved, congrats Will Merrill!

There are so many confounding factors here.
Our intuitions don't always match reality, and we really shouldn't make any decisions based on graphs like this.

Hi slack, it's me, the guy with 9 different channels on his slack app (with another quadzillion I can't reach with a shortcut).
You can stop targeting me with your ads.
Thank you.

All these secrets llm are hiding from us... It's almost as if they're sentient

LLM-as-judge is flawed, reason 371:

Share with us your data contamination insights, to improve our trust in evaluation.
huggingface.co/spaces/CONDA-W…

The most obnoxious citation type is when the authors didn't even read your paper.
It's even more frustrating when the authors didn't even read the freaking title 😡
I mean, common, at least read the title and abstract before you cite a paper...

🐣Too Long; Didn't Read: Most LLMs cite full documents for support, but they're too long so people don't actually read them.
In our new work 'Attribute First, then Generate”, we introduce fine-grained attributions, where LLMs are required to highlight only relevant information…

I am excited to be back at CLeaR and present this work today!

And no, 99% of the reviewers don't read whatever the new definitions mean for the latest conference

It is confusing because the definitions keep. changing. every. cycle. while being part of the same *ACL community.

Yesterday! Yanai Elazar came to @edinburghnlp @infated to talk about What’s In My Big Data from Allen Institute for AI

Giving a talk at EdinburghNLP tomorrow!
Come to hear about what makes it to text corpora these days and how it's affecting model behavior.

Thanks for having me, I had a lot of fun meeting everyone!

On my way to Munich to give a talk tomorrow at LMU.
Come to my talk if you're around!
DM if you wanna hang out

Shocking 😱😱😱