Phoebe Mulcaire @PhoebeNLP

Machine learning at @duolingo. Multilingual structured prediction and language modeling. Joined July 2018

Tweets

31
Followers

210
Following

64
Likes

332

Klinton Bicknell @klintonbicknell

a year ago

Duolingo AI is hiring new PhDs and PhD interns! Come define the future of AI-driven education! (New PhDs must graduate by Aug 2025, interns by Aug 2026) jobs.duolingo.com

8 80 316 47K 215

Matthew Finlayson @mattf1n

2 years ago

Wanna know gpt-3.5-turbo's embed size? We find a way to extract info from LLM APIs and estimate gpt-3.5-turbo’s embed size to be 4096. With the same trick we also develop 25x faster logprob extraction, audits for LLM APIs, and more! 📄 arxiv.org/abs/2403.09539 Here’s how 1/🧵

6 80 359 158K 169

Download Image

Alisa Liu @alisawuffles

2 years ago

LMs are increasingly large🐘 and proprietary🔒 — what if we could “tune”🔧 them without accessing their internal weights? Enter: proxy-tuning, which operates on only the *outputs* of LMs at decoding-time to achieve the effect of direct tuning! 📄: arxiv.org/abs/2401.08565 1/

3 83 370 63K 208

Download Image

Stephen Mayhew @mayhewsw

2 years ago

🚨 New Dataset Alert 🚨 I'm extremely excited to announce Universal NER v1, available now. It is gold-standard human annotations of 18 datasets covering 12 languages, based on Universal Dependencies texts. This is the first data release of the UNER project. 1/3

5 50 225 47K 95

Download Image

Phoebe Mulcaire @PhoebeNLP

2 years ago

at #ACL2023NLP ! I'll be presenting a poster at BEA on Thursday: work with @BenNaismithELT and @JillBurstein on discourse grading with GPT-4. come say hi :)

0 2 9 1K 0

Miles Turpin @milesaturpin

2 years ago

⚡️New paper!⚡️ It’s tempting to interpret chain-of-thought explanations as the LLM's process for solving a task. In this new work, we show that CoT explanations can systematically misrepresent the true reason for model predictions. arxiv.org/abs/2305.04388 🧵

13 113 505 146K 240

Download Image

Phoebe Mulcaire @PhoebeNLP

2 years ago

I'll be in Seattle from this Wednesday to next Tuesday—let me know if you'd like to meet up and say hi!

0 0 0 78 0

Maarten Sap (he/him) @MaartenSap

3 years ago

🌟Update🌟there's been a lot of debate of whether Theory of Mind (ToM) has emerged in new models (ChatGPT/GPT-3.5/GPT4), as people have reported qualitative/anecdotal evidence of good performance on these types of examples. TLDR; still no neural ToM... 🧵

Maarten Sap (he/him) @MaartenSap

3 years ago

18 131 702 0 336

7 53 205 98K 105

nostalgebraist @nostalgebraist

3 years ago

wait, what?? why do Bard AND ChatGPT *both* write an anodyne story about a young woman in idyllic "Willow Creek" at sundown??? (details: it's not deterministic, often you get a different town name, different phrasing, etc. broad strokes are similar though. gpt-4 does it too.)

7 10 85 27K 21

Download Image

Xiang Lisa Li @XiangLisaLi2

3 years ago

arxiv.org/abs/2210.15097 We propose contrastive decoding (CD), a more reliable search objective for text generation by contrasting LMs of different sizes. CD takes a large LM (expert LM e.g. OPT-13b) and a small LM (amateur LM e.g. OPT-125m) and maximizes their logprob difference

8 118 695 0 197

Download Image

John Thickstun @jwthickstun

3 years ago

In addition to proposing an interesting new truncation sampler, this paper draws a connection between truncation and n-gram smoothing that refined my understanding of the motivations for truncation sampling in general. This is a great read!

John Hewitt @johnhewtt

3 years ago

6 35 161 0 43

Download Image

0 2 7 0 2

Phoebe Mulcaire @PhoebeNLP

3 years ago

"improves over BERT on unseen scripts" struck me as really silly but it makes sense! fascinating work

Emanuele Bugliarello @ebugliarello

3 years ago

"improves over BERT on unseen scripts" struck me as really silly but it makes sense! fascinating work

25 179 999 0 285

Download Image

0 0 4 0 0

Phoebe Mulcaire @PhoebeNLP

3 years ago

I passed my defense! looking forward to starting at @duolingo this summer, working on ML for language learning.

3 2 170 0 1

Deyan Ginev @dginev

4 years ago

📢 Welcome to ar5iv.org Change the "X" in any arXiv article link to the "5" in ar5iv to get a modern HTML5 document. Thread: what is included, why now, and how we hope to merge back into arXiv. #OA #OpenScience #preprints 1/10

44 810 3K 0 431

Jungo Kasai 笠井淳吾 @jungokasai

4 years ago

We introduce 𝐁𝐢𝐝𝐢𝐦𝐞𝐧𝐬𝐢𝐨𝐧𝐚𝐥 𝐋𝐞𝐚𝐝𝐞𝐫𝐛𝐨𝐚𝐫𝐝𝐬 (𝐁𝐢𝐥𝐥𝐛𝐨𝐚𝐫𝐝𝐬), a generalization of leaderboards that simultaneously drives progress in natural language generation and its evaluation. arxiv.org/abs/2112.04139 nlp.cs.washington.edu/billboard/ 1/5

3 39 198 0 30

Download Image

Nikos Pappas @nik0spapp

5 years ago

How can we make language models be less data-hungry? We show that fully compositional word representations grounded on related words and definitions from an external lexicon can do that. Joint work w/ @PhoebeNLP,@nlpnoah at @emnlp2020 arxiv.org/abs/2009.11523 #NLProc (1/3)