Pan Wu
@panpanbb.bsky.social
Used to study neuroscience; Currently working as Machine Learning Engineer; Writing prompts everyday🥶🌀🌀
Reposted by Pan Wu
Cutting it a bit fine, but here’s my review of the year in neuroscience for 2024
The eighth of these, would you believe? We’ve got dark neurons, tiny monkeys, the most complete brain wiring diagram ever constructed, and much more…
Published on The Spike
Enjoy!
medium.com/the-spike/20...
The eighth of these, would you believe? We’ve got dark neurons, tiny monkeys, the most complete brain wiring diagram ever constructed, and much more…
Published on The Spike
Enjoy!
medium.com/the-spike/20...
2024: A Review of the Year in Neuroscience
Feeling a bit wired
medium.com
December 30, 2024 at 4:00 PM
Cutting it a bit fine, but here’s my review of the year in neuroscience for 2024
The eighth of these, would you believe? We’ve got dark neurons, tiny monkeys, the most complete brain wiring diagram ever constructed, and much more…
Published on The Spike
Enjoy!
medium.com/the-spike/20...
The eighth of these, would you believe? We’ve got dark neurons, tiny monkeys, the most complete brain wiring diagram ever constructed, and much more…
Published on The Spike
Enjoy!
medium.com/the-spike/20...
Reposted by Pan Wu
If you can’t wait for this release we have about 4 million books as part of Common Corpus. In multiple languages. huggingface.co/datasets/Ple... Easier subcollection browsing for cultural heritage monographs on here: huggingface.co/collections/...
December 12, 2024 at 9:30 PM
If you can’t wait for this release we have about 4 million books as part of Common Corpus. In multiple languages. huggingface.co/datasets/Ple... Easier subcollection browsing for cultural heritage monographs on here: huggingface.co/collections/...
Reposted by Pan Wu
Wrote this NodeJS script that can turn Starter Packs into Lists that you can follow and turn into custom Feeds.
Hoping this helps people build and recreate communities displaced from the Bad place.
Feel free to remix or modify, or build a public webapp!
gist.github.com/Makermed/dd0...
Hoping this helps people build and recreate communities displaced from the Bad place.
Feel free to remix or modify, or build a public webapp!
gist.github.com/Makermed/dd0...
A Scraper that turns Bluesky Starter Packs into Lists
A Scraper that turns Bluesky Starter Packs into Lists - starterpack_scraper.js
gist.github.com
November 26, 2024 at 4:49 AM
Wrote this NodeJS script that can turn Starter Packs into Lists that you can follow and turn into custom Feeds.
Hoping this helps people build and recreate communities displaced from the Bad place.
Feel free to remix or modify, or build a public webapp!
gist.github.com/Makermed/dd0...
Hoping this helps people build and recreate communities displaced from the Bad place.
Feel free to remix or modify, or build a public webapp!
gist.github.com/Makermed/dd0...
"Apple Intelligence is for the stupid ones," a journalist from CNET said in a recent video.
After watching Apple's ads, I guess I’m one of those "idiots"—because I really like the writing tools they offer.
November 26, 2024 at 6:02 AM
"Apple Intelligence is for the stupid ones," a journalist from CNET said in a recent video.
After watching Apple's ads, I guess I’m one of those "idiots"—because I really like the writing tools they offer.
feeds are down
a smiley face with red eyes and the words " it 's making me nervous " below it
ALT: a smiley face with red eyes and the words " it 's making me nervous " below it
media.tenor.com
November 26, 2024 at 2:59 AM
feeds are down
Can I sort the people I follow into different lists? Like, when I’m at work, I just want to see stuff from the tech list, but when I’m chilling and reading, I want updates from both the science and news lists.
November 24, 2024 at 11:44 AM
Can I sort the people I follow into different lists? Like, when I’m at work, I just want to see stuff from the tech list, but when I’m chilling and reading, I want updates from both the science and news lists.
Reposted by Pan Wu
Now that "scale is all we need" has predictably faltered...
1. Small AI models often perform better in context.
2. Obsession w bigness has bad consequences, from climate, to power concentration, to research capture.
Me + @gaelvaroquaux.bsky.social @sashamtl.bsky.social
arxiv.org/abs/2409.14160
1. Small AI models often perform better in context.
2. Obsession w bigness has bad consequences, from climate, to power concentration, to research capture.
Me + @gaelvaroquaux.bsky.social @sashamtl.bsky.social
arxiv.org/abs/2409.14160
Hype, Sustainability, and the Price of the Bigger-is-Better Paradigm in AI
With the growing attention and investment in recent AI approaches such as large language models, the narrative that the larger the AI system the more valuable, powerful and interesting it is is increa...
arxiv.org
November 17, 2024 at 11:16 AM
Now that "scale is all we need" has predictably faltered...
1. Small AI models often perform better in context.
2. Obsession w bigness has bad consequences, from climate, to power concentration, to research capture.
Me + @gaelvaroquaux.bsky.social @sashamtl.bsky.social
arxiv.org/abs/2409.14160
1. Small AI models often perform better in context.
2. Obsession w bigness has bad consequences, from climate, to power concentration, to research capture.
Me + @gaelvaroquaux.bsky.social @sashamtl.bsky.social
arxiv.org/abs/2409.14160
Reposted by Pan Wu
Great to see Anthropic highlighting the importance of uncertainty quantification (UQ) in AI.
Having spent my PhD and postdoc developing scalable UQ methods for real-world, high-dimensional problems, I’m currently on the academic job market—feel free to reach out!
Having spent my PhD and postdoc developing scalable UQ methods for real-world, high-dimensional problems, I’m currently on the academic job market—feel free to reach out!
A statistical approach to model evaluations
A research paper from Anthropic on how to apply statistics to improve language model evaluations
www.anthropic.com
November 21, 2024 at 8:04 PM
Great to see Anthropic highlighting the importance of uncertainty quantification (UQ) in AI.
Having spent my PhD and postdoc developing scalable UQ methods for real-world, high-dimensional problems, I’m currently on the academic job market—feel free to reach out!
Having spent my PhD and postdoc developing scalable UQ methods for real-world, high-dimensional problems, I’m currently on the academic job market—feel free to reach out!
I like Bluesky more. Threads is like, “Oh, you’re in Hong Kong ? Stay local ! No tech/science posts from outside !” It feels like Facebook—just people near me. Twitter (pre-X) showed me the world. And Bluesky brings that feeling back. Hello world again 🙋🏻♀️🙋🏻♀️
November 22, 2024 at 4:33 AM
I like Bluesky more. Threads is like, “Oh, you’re in Hong Kong ? Stay local ! No tech/science posts from outside !” It feels like Facebook—just people near me. Twitter (pre-X) showed me the world. And Bluesky brings that feeling back. Hello world again 🙋🏻♀️🙋🏻♀️
Reposted by Pan Wu
How do LLMs learn to reason from data? Are they ~retrieving the answers from parametric knowledge🦜? In our new preprint, we look at the pretraining data and find evidence against this:
Procedural knowledge in pretraining drives LLM reasoning ⚙️🔢
🧵⬇️
Procedural knowledge in pretraining drives LLM reasoning ⚙️🔢
🧵⬇️
November 20, 2024 at 4:35 PM
How do LLMs learn to reason from data? Are they ~retrieving the answers from parametric knowledge🦜? In our new preprint, we look at the pretraining data and find evidence against this:
Procedural knowledge in pretraining drives LLM reasoning ⚙️🔢
🧵⬇️
Procedural knowledge in pretraining drives LLM reasoning ⚙️🔢
🧵⬇️
Reposted by Pan Wu
Men are floundering at school and in the workplace. Some conservatives blame a crisis of masculinity, but the problems—and their solutions—are far more complex.
What’s the Matter with Men?
They’re floundering at school and in the workplace. Some conservatives blame a crisis of masculinity, but the problems—and their solutions—are far more complex.
www.newyorker.com
November 22, 2024 at 12:23 AM
Men are floundering at school and in the workplace. Some conservatives blame a crisis of masculinity, but the problems—and their solutions—are far more complex.