Alex Wettig @_awettig

phd @Princeton agents @cursor_ai cs.princeton.edu/~awettig/ Joined July 2022

Tweets

189
Followers

2K
Following

593
Likes

2K

Cursor @cursor_ai

8 hours ago

Cursor can now control your browser. Agent can take screenshots, improve UI, and debug client issues. Try our early preview with Sonnet 4.5.

121 227 3K 280K 1K

Download Video

We evaluated Anthropic's Sonnet 4.5 with our minimal agent. New record on SWE-bench verified: 70.6%! Same price/token as Sonnet 4, but takes more steps, ending up being more expensive. Cost analysis details & link to full trajectories in 🧵

3 6 37 18K 15

Download Image

Charlie Snell @sea_snell

4 weeks ago

yolo run summer is over scaling laws fall has arrived

Preet Sojitra @Preet_Sojitra03

7 months ago

yolo run summer is over scaling laws fall has arrived

2 0 26 23K 1

Download Image

1 1 63 8K 4

Jessy Lin @realJessyLin

a month ago

🔍 How do we teach an LLM to 𝘮𝘢𝘴𝘵𝘦𝘳 a body of knowledge? In new work with @AIatMeta, we propose Active Reading 📙: a way for models to teach themselves new things by self-studying their training data. Results: * 𝟔𝟔% on SimpleQA w/ an 8B model by studying the wikipedia…

15 157 1K 129K 1K

Download Image

Stuart Sul @stuart_sul

a month ago

MoE layers can be really slow. When training our coding models @cursor_ai, they ate up 27–53% of training time. So we completely rebuilt it at the kernel level and transitioned to MXFP8. The result: 3.5x faster MoE layer and 1.5x end-to-end training speedup. We believe our…

30 105 875 142K 427

Download Image

Alex Wettig @_awettig

3 months ago

Presenting two posters at ICML over the next two days: - Both at 11am - 1:30pm - Both about how to improve pre-training with domains - Both at stall # E-2600 in East Exhibition Hall A-B (!) Tomorrow: WebOrganizer w/ @soldni & @kylelostat Thursday: MeCo by @gaotianyu1350

1 9 49 11K 8

Download Image

Albert Gu @_albertgu

3 months ago

Tokenization is just a special case of "chunking" - building low-level data into high-level abstractions - which is in turn fundamental to intelligence. Our new architecture, which enables hierarchical *dynamic chunking*, is not only tokenizer-free, but simply scales better.

Sukjun (June) Hwang @sukjun_hwang

3 months ago

98 744 5K 745K 4K

Download Gif

61 194 1K 203K 770

Download Image

Anthropic @AnthropicAI

3 months ago

Anthropic staff realized they could ask Claude to buy things that weren’t just food & drink. After someone randomly decided to ask it to order a tungsten cube, Claude ended up with an inventory full of (as it put it) “specialty metal items” that it ended up selling at a loss.

65 211 4K 845K 326

Download Image

Alex Wettig @_awettig

3 months ago

New paper cutting through the thicket of KV cache eviction methods!

Adithya Bhaskar @AdithyaNLP

3 months ago

New paper cutting through the thicket of KV cache eviction methods!

2 36 233 31K 194

Download Image

0 1 17 1K 1

Alex Zhang @a1zhang

4 months ago

Can GPT, Claude, and Gemini play video games like Zelda, Civ, and Doom II? 𝗩𝗶𝗱𝗲𝗼𝗚𝗮𝗺𝗲𝗕𝗲𝗻𝗰𝗵 evaluates VLMs on Game Boy & MS-DOS games given only raw screen input, just like how a human would play. The best model (Gemini) completes just 0.48% of the benchmark! 🧵👇

23 77 547 123K 235

Download Video

Aman Sanger @amanrsanger

4 months ago

Claude Sonnet 4 is much better at codebase understanding. Paired with recent improvements in Cursor, it's SOTA on large codebases

32 43 859 93K 112

Download Image

Kilian Lieret @KLieret

4 months ago

Massive gains with Sonnet 4 on SWE-agent: Single-attempt pass@1 rises to 69% on SWE-bench Verified! Sonnet 4 iterates longer (making it slightly more expensive) but almost never gets stuck. Localization ability appears unchanged, but quality of edits improves.

4 14 85 9K 35

Download Image

Ofir Press @OfirPress

4 months ago

Great results from the Claude team- the 80% result is pass@1!! They ran the model in parallel multiple times and had an LM judge pick the best patch to submit.

5 7 120 15K 14

Download Image

Alex Wettig @_awettig

5 months ago

Big arrow time! We can make huge progress on open-source SWE agents by scaling up the creation of virtual coding environments 🚀

John Yang @jyangballin

5 months ago

Big arrow time! We can make huge progress on open-source SWE agents by scaling up the creation of virtual coding environments 🚀

24 139 655 100K 377

Download Image

0 4 17 3K 4

Cursor @cursor_ai

5 months ago

Cursor is now free for students. Enjoy!

2K 4K 41K 11.2M 15K

Xindi Wu @cindy_x_wu

5 months ago

Introducing COMPACT: COMPositional Atomic-to-complex Visual Capability Tuning, a data-efficient approach to improve multimodal models on complex visual tasks without scaling data volume. 📦 arxiv.org/abs/2504.21850 1/10

6 48 161 53K 76

Download Image

John Yang @jyangballin

5 months ago

@ weekend warriors - DM me a GitHub repo that you like / maintain, and I'll train you a 7B coding agent that's an expert for that repo. Main constraints - it's predominantly Python, and has a testing suite w/ good coverage. (example of good repo = sympy, pandas, sqlfluff)

19 8 115 17K 66

Jacob Springer @jacspringer

6 months ago

Training with more data = better LLMs, right? 🚨 False! Scaling language models by adding more pre-training data can decrease your performance after post-training! Introducing "catastrophic overtraining." 🥁🧵+arXiv 👇 1/9

17 185 821 162K 655

Download Image

Alisa Liu @alisawuffles

6 months ago

We created SuperBPE🚀, a *superword* tokenizer that includes tokens spanning multiple words. When pretraining at 8B scale, SuperBPE models consistently outperform the BPE baseline on 30 downstream tasks (+8% MMLU), while also being 27% more efficient at inference time.🧵

94 326 3K 364K 1K

Download Image

Logan Engstrom @logan_engstrom

6 months ago

Want state-of-the-art data curation, data poisoning & more? Just do gradient descent! w/ @andrew_ilyas Ben Chen @axel_s_feldmann @wsmoses @aleks_madry: we show how to optimize final model loss wrt any continuous variable. Key idea: Metagradients (grads through model training)