Jackson Petty
@jacksonpetty.org
📤 210
📥 254
📝 313
the passionate shepherd, to his love • ἀρετῇ • מנא הני מילי
pinned post!
Pingali and Bilardi (2015) just get me
over 1 year ago
1
6
1
reposted by
Jackson Petty
Ariel Edwards-Levy
6 months ago
"there's a new serif in town"
add a skeleton here at some point
1008
14731
2752
reposted by
Jackson Petty
Greg Durrett
6 months ago
📢 Postdoc position 📢 I’m recruiting a postdoc for my lab at NYU! Topics include LM reasoning, creativity, limitations of scaling, AI for science, & more! Apply by Feb 1. (Different from NYU Faculty Fellows, which are also great but less connected to my lab.) Link in 🧵
2
21
13
reposted by
Jackson Petty
NYU Center for Data Science
9 months ago
Linguistics PhD student
@jacksonpetty.org
finds LLMs "quiet-quit" when instructions get long, switching from reasoning to guesswork. With CDS'
@tallinzen.bsky.social
,
@shauli.bsky.social
,
@lambdaviking.bsky.social
,
@michahu.bsky.social
, and Wentao Wang.
nyudatascience.medium.com/llms-switch-...
loading . . .
LLMs Switch to Guesswork Once Instructions Get Long
LLMs abandon reasoning for guesswork when instructions get long, new work from Linguistics PhD student Jackson Petty & CDS shows.
https://nyudatascience.medium.com/llms-switch-to-guesswork-once-instructions-get-long-b74dbf6f05ae
0
7
2
you’re telling me a star spangled this banner??
11 months ago
0
1
0
reposted by
Jackson Petty
Tal Linzen
12 months ago
I'm hiring at least one post-doc! We're interested in creating language models that process language more like humans than mainstream LLMs do, through architectural modifications and interpretability-style steering. Express interest here:
docs.google.com/forms/d/e/1F...
loading . . .
NYU LLM + cognitive science post-doc interest form
Tal Linzen's group at NYU is hiring a post-doc! We're interested in creating language models that process language more like humans than mainstream LLMs do, through architectural modifications and int...
https://docs.google.com/forms/d/e/1FAIpQLSc5IwTU43CWVjQYsWbvPkDFH7dFKglqRfPdWRJSvCbYuxlv-A/viewform
2
42
22
How well can LLMs understand tasks with complex sets of instructions? We investigate through the lens of RELIC: REcognizing (formal) Languages In-Context, finding a significant overhang between what LLMs are able to do theoretically and how well they put this into practice.
12 months ago
1
5
2
[guy who only watched the pilot episode of Caprica voice] Nice!
add a skeleton here at some point
over 1 year ago
0
1
0
reposted by
Jackson Petty
Kyle Mahowald
over 1 year ago
Looking forward to speaking tomorrow (Tues am) in this Simons workshop in Berkeley
simons.berkeley.edu/workshops/ll...
. Will talk about some empirical work and also share some takes from this recent preprint from me and
@futrell.bsky.social
arxiv.org/abs/2501.17047
loading . . .
LLMs, Cognitive Science, Linguistics, and Neuroscience
At a conceptual level, LLMs profoundly change the landscape for theories of human language, of the brain and computation, and of the nature of human intelligence. In linguistics, they provide a new wa...
https://simons.berkeley.edu/workshops/llms-cognitive-science-linguistics-neuroscience
0
10
2
@foodnoms.com
Can I submit a feature request on bsky? It would be amazing if we could tie Goals to bodyweight, as read in from Apple Health data. Many recommendations for daily protein intake, for instance, are given in grams/kg of bodyweight.
over 1 year ago
1
1
0
reposted by
Jackson Petty
Eighth night // Dedication of the House (Shimen Frug)
over 1 year ago
0
0
1
reposted by
Jackson Petty
Seventh night // Dedication of the House (Shimen Frug)
over 1 year ago
1
0
1
reposted by
Jackson Petty
Happy Hanukkah
over 1 year ago
1
1
1
reposted by
Jackson Petty
shabbat shalom
over 1 year ago
1
0
1
reposted by
Jackson Petty
בימים ההם בזמן הזה
over 1 year ago
1
0
1
Accidentally bought tickets to the Sing Along version of Nosferatu 😑
over 1 year ago
0
0
0
Happy Hanukkah
over 1 year ago
1
3
0
reposted by
Jackson Petty
Later on, we'll compile, As we train on The Pile, To autoregress The next words we'll guess Grokking in the latent feature land
over 1 year ago
1
1
1
I think 2m43s is a new personal record
over 1 year ago
0
1
0
o3, o3 in the loop, list the steps in making soup
over 1 year ago
0
0
0
reposted by
Jackson Petty
over 1 year ago
Well one couldn't help but become mildly sensitive to the context over there!
1
1
1
“You’re telling me a Tree Adjoined this Grammar??” —
@bob-frank.bsky.social
over 1 year ago
1
2
0
Claude understood the assignment (a bit too well)
add a skeleton here at some point
over 1 year ago
0
0
0
URL verification works fine. The bigger issue is that people are reluctant to use a URL as their handle without the ability to reserve the *.bsky.social version and have it redirect
add a skeleton here at some point
over 1 year ago
1
0
0
Of all sad words of tongue or pen, the saddest are these:
over 1 year ago
1
3
0
reposted by
Jackson Petty
Dr. JD
over 1 year ago
People who regularly use emojis, please choose the following sentence which is least acceptable. 1️⃣
🫵 said that 🫵 would help me!
2️⃣
🫵 said that you would help me!
3️⃣
You said that 🫵 would help me!
4️⃣
I have no judgments on this
📊 Show results
0
3
3
girlfriend tax implies the existence of the girlfriend laugher curve, which is the theoretical relationship between how bad my jokes are and how much she laughs at them
add a skeleton here at some point
over 1 year ago
0
0
0
girlfriend tax implies the existence of girlfriend industrial policy
over 1 year ago
0
1
1
home economics, but I repeat myself
over 1 year ago
0
1
0
reposted by
Jackson Petty
Barbara Partee ‘Noun Phrase Interpretation and Type-shifting Principles’ (1987) // Sir Harris Nicolas ‘Aigncourt banner of the Trinity’ (1833)
about 3 years ago
0
2
1
reposted by
Jackson Petty
about 3 years ago
0
2
1
1) what
over 1 year ago
0
1
0
polymarket for which track my train to boston is leaving penn station on
over 1 year ago
1
0
0
so apparently the Ancient Greek verb for “to hurt” is just “aaaaaooooow”
over 1 year ago
1
19
5
You should go work with Isabel
add a skeleton here at some point
over 1 year ago
0
0
0
In case it’s of use to other academics, I put together an overleaf project which shows how to use different styles for in-text citations and bibliographies (eg, author-year in text vs numeric in bib). Works in BibLaTeX and hackily in natbib. Link:
www.overleaf.com/read/tkxbfty...
over 1 year ago
1
2
0
Pingali and Bilardi (2015) just get me
over 1 year ago
1
6
1
A highlight of EMNLP: finding a fellow Heliotrope enthusiast in
@venkatasg.net
over 1 year ago
1
1
0
Scenes from Miami
over 1 year ago
0
2
0
over 1 year ago
1
0
0
reposted by
Jackson Petty
Tom McCoy
over 1 year ago
🤖🧠 I'll be considering applications for postdocs & PhD students to start at Yale in Fall 2025! If you are interested in the intersection of linguistics, cognitive science, and AI, I encourage you to apply! Postdoc link:
rtmccoy.com/prospective_...
PhD link:
rtmccoy.com/prospective_...
0
14
5
November in New Haven
over 1 year ago
0
0
0
“Something is ROT-10 in the state of Denmark” slow clap,
@rtommccoy.bsky.social
, slow clap
over 1 year ago
0
1
0
if you’re at EMNLP and wanna brainstorm ideas around: - LLMs for ancient languages - code & formal languages for LLMs - applying (😱) math to better understand LLMs we should chat this week!
over 1 year ago
0
0
1
Are you coming to
#EMNLP2024
? Come say hi! I’ll be presenting this work at BlackboxNLP on Friday, and will be around all week to chat! DM to meet up, let’s grab lunch!
over 1 year ago
0
7
0
Our paper on how model depth impacts compositional generalization was accepted to
#NAACL2024
🇲🇽 A new version is up on arXiv! We've added a discussion of how making models deeper makes them slower, without adding much benefit for compositional tasks
arxiv.org/abs/2310.19956
add a skeleton here at some point
about 2 years ago
0
4
2
How well do LLMs generalize to novel distributions? Our (Michael Wilson Bob Frank) new TACL / EMNLP 2023 paper uses argument structure to test this, finding that LLMs succeed within known contexts but fail to do so between appropriately related ones. 🧵
jacksonpetty.org/argstruct
loading . . .
How Abstract Is Linguistic Generalization in Large Language Models? Experiments with Argument Struct...
Language models are typically evaluated on their success at predicting the distribution of specific words in specific contexts. Yet linguistic knowledge also encodes relationships between contexts, al...
https://jacksonpetty.org/argstruct
over 2 years ago
1
2
0
Do deep transformer LMs generalize better? In a new preprint we (Sjoerd van Steenkiste, Ishita Dasgupta, Fei Sha, Dan Garrette, &
@tallinzen.bsky.social
) control for parameter count to show how depth helps models on compositional generalization tasks, but diminishingly so 🧵
jacksonpetty.org/depth
loading . . .
The Impact of Depth and Width on Transformer Language Model Generalization
To process novel sentences, language models (LMs) must generalize compositionally -- combine familiar elements in new ways. What aspects of a model's structure promote compositional generalization? Fo...
https://jacksonpetty.org/depth
over 2 years ago
1
10
6
Load more
feeds!
log in