Zhu

hello world
@penn NLP PhD
🦋

[Zhu @zhuexe]

oh hey that's me

[elvis @omarsar0]

Overhearing LLM Agents

Having AI agents proactively suggest and provide additional context can improve many workflows.

This is an exciting direction on AI agents that I've been exploring for the last couple of months, mostly for note-taking and research.

[Zhu @zhuexe]

#ACL2024 Come talk to me about software stuff and NLP, or don't, I'm not your advisor

[Chris Stokel-Walker @stokel]

Good morning. Stay in bed today. Early reports suggest Crowdstrike Falcon - a computer threat checker used by lots (and lots and lots) of businesses pushed out an update that might have broken a lot of computers. Airlines, businesses etc affected

[Liam Dugan @LiamDugan_]

🚨New Paper🚨: Are AI text detectors *really* as good as they claim? (#ACL2024)

We release RAID—The largest & most challenging detection benchmark with 6M+ outputs from 11 LLMs, 8 domains, 4 decoding strategies, and 11 adv attacks

https://t.co/mPlNymdchJ
https://t.co/VR0mMvqXRm

[Zhu @zhuexe]

Today is Friday.

[Republic of Kiribati @KiribatiGov]

Fine. You're not invited

[Jascha Sohl-Dickstein @jaschasd]

Have you ever done a dense grid search over neural network hyperparameters? Like a *really dense* grid search? It looks like this (!!). Blueish colors correspond to hyperparameters for which training converges, redish colors to hyperparameters for which training diverges.

[Michael Irving @MikeIrvo]

The audio log you find in a trashed office in a sci-fi horror game:

[Conference on Language Modeling @COLM_conf]

We are pleased to announce that the first Conference on Language Modeling will be held at the University of Pennsylvania in Philadelphia at the Zellerbach Theatre.

Thanks so much to UPenn CS as well as Mark Yatskar and Zachary Ives for facilitating the amazing venue.

[Zhu @zhuexe]

twitter doesn't like it when you say the names of not-twitter things nowadays right

anyway i'm https://t.co/8FqGDQkMRq on ceruleanatmosphere

[Zhu @zhuexe]

wat

[Prithviraj (Raj) Ammanabrolu @rajammanabrolu]

The PEARLS Lab at @UCSD_CSE is now open for business! I'm recruiting Fall 24 PhD students in all things interactive and grounded AI, RL, and NLP!! Join us in the land of 🏖️ beach (🧋pearl tea included). Apply by Dec 20. Please help spread the word!

More: https://t.co/nPZT3KbuUC

[Prithviraj (Raj) Ammanabrolu @rajammanabrolu]

Every single closed source model rumor leak on this site

[Liam Dugan @LiamDugan_]

Excited to announce that Kani has officially been accepted and will appear at the NLP Open Source Software Workshop at #EMNLP2023!

See you all in Singapore :)

[Zhu @zhuexe]

how do i use twitter

[Liam Dugan @LiamDugan_]

✨New Paper✨: We release Kani 🦀 a highly-hackable open-source library for building LM apps with tool usage (e.g. plugins)

Kani lets you easily write LM-callable functions in pure Python w/ robust type checks + model retry

https://t.co/CfqvZTboFo
https://t.co/NmsjD5fI1Q

[Zhu @zhuexe]

@CamBamJamFam @NicoleSHsing long time no see

[Shal Marriott @ShalMarriott]

Academics when they realize it’s August.

[Zhu @zhuexe]

We're at ACL 2023! Come check out our poster at Session 7 (Wed 11-12:30) in the Frontenac Ballroom!

[Zhu @zhuexe]

✨New Paper📷: FIREBALL! is the largest dataset containing structured game states from real Dungeons and Dragons games, capturing 8 million utterances and 1.3 million game states. https://t.co/sKGAJLjh0f