Shivam Raval
@sraval.bsky.social
Physics, Visualization and AI PhD @ Harvard | Embedding visualization and LLM interpretability | Love pretty visuals, math, physics and pets | Currently into manifolds
Wanna meet and chat? Book a meeting here: https://zcal.co/shivam-raval
Wanna meet and chat? Book a meeting here: https://zcal.co/shivam-raval
Reposted by Shivam Raval
This map shows the hour of sunrise globally through the year. It reveals time zones following national and, sometimes, regional boundaries, and slicing through the oceans.
May 12, 2025 at 1:25 PM
This map shows the hour of sunrise globally through the year. It reveals time zones following national and, sometimes, regional boundaries, and slicing through the oceans.
Update on the VIS+AI meetup: I'm a speaker now!
April 3, 2025 at 7:37 PM
Update on the VIS+AI meetup: I'm a speaker now!
Reposted by Shivam Raval
Can we understand the mechanisms of a frontier AI model?
📝 Blog post: www.anthropic.com/research/tra...
🧪 "Biology" paper: transformer-circuits.pub/2025/attribu...
⚙️ Methods paper: transformer-circuits.pub/2025/attribu...
Featuring basic multi-step reasoning, planning, introspection and more!
📝 Blog post: www.anthropic.com/research/tra...
🧪 "Biology" paper: transformer-circuits.pub/2025/attribu...
⚙️ Methods paper: transformer-circuits.pub/2025/attribu...
Featuring basic multi-step reasoning, planning, introspection and more!
On the Biology of a Large Language Model
transformer-circuits.pub
March 27, 2025 at 6:18 PM
Can we understand the mechanisms of a frontier AI model?
📝 Blog post: www.anthropic.com/research/tra...
🧪 "Biology" paper: transformer-circuits.pub/2025/attribu...
⚙️ Methods paper: transformer-circuits.pub/2025/attribu...
Featuring basic multi-step reasoning, planning, introspection and more!
📝 Blog post: www.anthropic.com/research/tra...
🧪 "Biology" paper: transformer-circuits.pub/2025/attribu...
⚙️ Methods paper: transformer-circuits.pub/2025/attribu...
Featuring basic multi-step reasoning, planning, introspection and more!
📭 Announcing two new blog posts!
One discusses a review ethics violation from last year. ieeevis.org/blog/vis-202...
The other describes ongoing efforts to revise the organizational structure of VIS. ieeevis.org/blog/vis-202...
One discusses a review ethics violation from last year. ieeevis.org/blog/vis-202...
The other describes ongoing efforts to revise the organizational structure of VIS. ieeevis.org/blog/vis-202...
ieeevis.org
March 21, 2025 at 8:41 PM
Join us for our first Vis+AI meetup on April 3rd at Northeastern University, a meetup to gather people interested in the intersection of Data Visualization and Artificial Intelligence. Sign up as soon as possible! We have a limited number of spots. lnkd.in/e8whS6v2.
March 18, 2025 at 9:47 PM
Join us for our first Vis+AI meetup on April 3rd at Northeastern University, a meetup to gather people interested in the intersection of Data Visualization and Artificial Intelligence. Sign up as soon as possible! We have a limited number of spots. lnkd.in/e8whS6v2.
Reposted by Shivam Raval
The wind map at hint.fm/wind/ has been running since 2012, relying on weather data from NOAA. We added a notice like this today. Thanks to @cambecc.bsky.social for the inspiration.
March 3, 2025 at 1:56 AM
The wind map at hint.fm/wind/ has been running since 2012, relying on weather data from NOAA. We added a notice like this today. Thanks to @cambecc.bsky.social for the inspiration.
Reposted by Shivam Raval
Great thread describing the new ARBOR open interpretability project, which has some fascinating projects already. Take a look!
ARBOR aims to accelerate the internal investigation of the new class of AI "reasoning" models.
See the ARBOR discussion board for a thread for each project underway.
github.com/ArborProjec...
See the ARBOR discussion board for a thread for each project underway.
github.com/ArborProjec...
February 20, 2025 at 10:49 PM
Great thread describing the new ARBOR open interpretability project, which has some fascinating projects already. Take a look!
Reposted by Shivam Raval
Today we're launching a multi-lab open collaboration, the ARBOR project, to accelerate AI interpretability research for reasoning models. Please join us!
github.com/ARBORproject...
(ARBOR = Analysis of Reasoning Behavior through Open Research)
github.com/ARBORproject...
(ARBOR = Analysis of Reasoning Behavior through Open Research)
GitHub - ARBORproject/arborproject.github.io
Contribute to ARBORproject/arborproject.github.io development by creating an account on GitHub.
github.com
February 20, 2025 at 7:55 PM
Today we're launching a multi-lab open collaboration, the ARBOR project, to accelerate AI interpretability research for reasoning models. Please join us!
github.com/ARBORproject...
(ARBOR = Analysis of Reasoning Behavior through Open Research)
github.com/ARBORproject...
(ARBOR = Analysis of Reasoning Behavior through Open Research)
Reposted by Shivam Raval
DeepSeek R1 shows how important it is to be studying the internals of reasoning models. Try our code: Here @canrager.bsky.social shows a method for auditing AI bias by probing the internal monologue.
dsthoughts.baulab.info
I'd be interested in your thoughts.
dsthoughts.baulab.info
I'd be interested in your thoughts.
dsthoughts.baulab
January 31, 2025 at 2:30 PM
DeepSeek R1 shows how important it is to be studying the internals of reasoning models. Try our code: Here @canrager.bsky.social shows a method for auditing AI bias by probing the internal monologue.
dsthoughts.baulab.info
I'd be interested in your thoughts.
dsthoughts.baulab.info
I'd be interested in your thoughts.
Reposted by Shivam Raval
In 1897, Alfred G. Mayer created his butterfly wing projections, an attempt to gain new insights into natural patterns and laws. Vertical blocks denote individual wings, distorted and stretched mathematically to fill a tidy rectangular space. More here: publicdomainreview.org/collection/m...
January 14, 2025 at 9:23 PM
In 1897, Alfred G. Mayer created his butterfly wing projections, an attempt to gain new insights into natural patterns and laws. Vertical blocks denote individual wings, distorted and stretched mathematically to fill a tidy rectangular space. More here: publicdomainreview.org/collection/m...
Reposted by Shivam Raval
DeepSeek is a side project 🔥
January 24, 2025 at 8:01 AM
DeepSeek is a side project 🔥
Reposted by Shivam Raval
Reposted by Shivam Raval
Tailscan website now uses v4!
Also updated the Tailwind CSS color palette cheat sheet 👀 added a button to see the old v3 and new v4 color Tailwind color palette.
#buildinpublic
Also updated the Tailwind CSS color palette cheat sheet 👀 added a button to see the old v3 and new v4 color Tailwind color palette.
#buildinpublic
January 23, 2025 at 3:18 PM
Tailscan website now uses v4!
Also updated the Tailwind CSS color palette cheat sheet 👀 added a button to see the old v3 and new v4 color Tailwind color palette.
#buildinpublic
Also updated the Tailwind CSS color palette cheat sheet 👀 added a button to see the old v3 and new v4 color Tailwind color palette.
#buildinpublic
Reposted by Shivam Raval
I'm teaching my first course! A seminar on "Machine Behavior."
Readings are a mix of NLP, CSS-y, and ML work on how machines (focus LLMs) "behave" within sociotechnical systems and on how they can be used to study human behavior.
Syllabus: manoelhortaribeiro.github.io/teaching/spr...
Readings are a mix of NLP, CSS-y, and ML work on how machines (focus LLMs) "behave" within sociotechnical systems and on how they can be used to study human behavior.
Syllabus: manoelhortaribeiro.github.io/teaching/spr...
January 20, 2025 at 3:04 PM
I'm teaching my first course! A seminar on "Machine Behavior."
Readings are a mix of NLP, CSS-y, and ML work on how machines (focus LLMs) "behave" within sociotechnical systems and on how they can be used to study human behavior.
Syllabus: manoelhortaribeiro.github.io/teaching/spr...
Readings are a mix of NLP, CSS-y, and ML work on how machines (focus LLMs) "behave" within sociotechnical systems and on how they can be used to study human behavior.
Syllabus: manoelhortaribeiro.github.io/teaching/spr...
Reposted by Shivam Raval
1) In the narrow area of prompt generation techniques LLMs can generate ideas rated as more novel and exciting. They are sometimes less feasible. Out of 4000 ideas generated, only 200 were potentially unique. arxiv.org/abs/2409.04109
Can LLMs Generate Novel Research Ideas? A Large-Scale Human Study with 100+ NLP Researchers
Recent advancements in large language models (LLMs) have sparked optimism about their potential to accelerate scientific discovery, with a growing number of works proposing research agents that autono...
arxiv.org
January 20, 2025 at 3:10 PM
1) In the narrow area of prompt generation techniques LLMs can generate ideas rated as more novel and exciting. They are sometimes less feasible. Out of 4000 ideas generated, only 200 were potentially unique. arxiv.org/abs/2409.04109
Reposted by Shivam Raval
hahahahah there were actually two technical reports for RL reasoning models today, kimi 1.5 also has good stuff on reward shaping + RL infra
kimi 1.5 report: https://buff.ly/4jqgCOa
kimi 1.5 report: https://buff.ly/4jqgCOa
January 20, 2025 at 3:55 PM
hahahahah there were actually two technical reports for RL reasoning models today, kimi 1.5 also has good stuff on reward shaping + RL infra
kimi 1.5 report: https://buff.ly/4jqgCOa
kimi 1.5 report: https://buff.ly/4jqgCOa
Reposted by Shivam Raval
Dimensionality reduction beyond neural subspaces with slice tensor component analysis
www.nature.com/articles/s41...
www.nature.com/articles/s41...
Dimensionality reduction beyond neural subspaces with slice tensor component analysis - Nature Neuroscience
Neural activity does not always lie in a low-dimensional subspace. The authors extend this classic view to show that task-relevant information is distributed across multiple covariability classes and ...
www.nature.com
October 1, 2024 at 12:44 PM
Dimensionality reduction beyond neural subspaces with slice tensor component analysis
www.nature.com/articles/s41...
www.nature.com/articles/s41...
Reposted by Shivam Raval
What a beauty! This is comet C/2024 G3 (ATLAS) passing through the field of view of the LASCO C3 coronagraph.
It wasn't for certain whether it would survive it's closest approach to the sun on January 13th, but it did and delivered us a spectacular show!
#comet #C2024G3 🔭
It wasn't for certain whether it would survive it's closest approach to the sun on January 13th, but it did and delivered us a spectacular show!
#comet #C2024G3 🔭
January 16, 2025 at 9:14 PM
Reposted by Shivam Raval
To wrap your head around agents and think through the ethics, our Society and Ethics team put together a great resource - 👏 @mmitchell.bsky.social @sashamtl.bsky.social @evijit.io @giadapistilli.com
huggingface.co/blog/ethics-...
huggingface.co/blog/ethics-...
AI Agents Are Here. What Now?
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
huggingface.co
January 14, 2025 at 8:27 PM
To wrap your head around agents and think through the ethics, our Society and Ethics team put together a great resource - 👏 @mmitchell.bsky.social @sashamtl.bsky.social @evijit.io @giadapistilli.com
huggingface.co/blog/ethics-...
huggingface.co/blog/ethics-...
Reposted by Shivam Raval
Pie and donut charts get a bad rep, but they work well if used for the right data and tasks. Read about what the science has to say about them in our new blog post: https://buff.ly/3DURbnS
January 14, 2025 at 10:00 PM
Pie and donut charts get a bad rep, but they work well if used for the right data and tasks. Read about what the science has to say about them in our new blog post: https://buff.ly/3DURbnS
Reposted by Shivam Raval
*Deep Learning Through A Telescoping Lens*
by @alanjeffares.bsky.social @aliciacurth.bsky.social
Shows that tracking 1st-order approximations to the training dynamics provides insights into many phenomena (e.g., double descent, grokking).
arxiv.org/abs/2411.00247
by @alanjeffares.bsky.social @aliciacurth.bsky.social
Shows that tracking 1st-order approximations to the training dynamics provides insights into many phenomena (e.g., double descent, grokking).
arxiv.org/abs/2411.00247
January 14, 2025 at 10:43 AM
*Deep Learning Through A Telescoping Lens*
by @alanjeffares.bsky.social @aliciacurth.bsky.social
Shows that tracking 1st-order approximations to the training dynamics provides insights into many phenomena (e.g., double descent, grokking).
arxiv.org/abs/2411.00247
by @alanjeffares.bsky.social @aliciacurth.bsky.social
Shows that tracking 1st-order approximations to the training dynamics provides insights into many phenomena (e.g., double descent, grokking).
arxiv.org/abs/2411.00247
Reposted by Shivam Raval
New paper <3
Interested in inference-time scaling? In-context Learning? Mech Interp?
LMs can solve novel in-context tasks, with sufficient examples (longer contexts). Why? Bc they dynamically form *in-context representations*!
1/N
Interested in inference-time scaling? In-context Learning? Mech Interp?
LMs can solve novel in-context tasks, with sufficient examples (longer contexts). Why? Bc they dynamically form *in-context representations*!
1/N
January 5, 2025 at 3:49 PM
New paper <3
Interested in inference-time scaling? In-context Learning? Mech Interp?
LMs can solve novel in-context tasks, with sufficient examples (longer contexts). Why? Bc they dynamically form *in-context representations*!
1/N
Interested in inference-time scaling? In-context Learning? Mech Interp?
LMs can solve novel in-context tasks, with sufficient examples (longer contexts). Why? Bc they dynamically form *in-context representations*!
1/N
Reposted by Shivam Raval
I've started a Research Integrity Feed populated by hashtags below & choice users. For the #SciPub / #AcademicPublishing sleuths 📊🔍👀
Plus it's got a cute furry mascot! 😘
bsky.app/profile/did:...
#ResearchIntegrity
#PredatoryPublisher
#PredatoryPublishing
#EditorialIndependence
#SciRetraction
Plus it's got a cute furry mascot! 😘
bsky.app/profile/did:...
#ResearchIntegrity
#PredatoryPublisher
#PredatoryPublishing
#EditorialIndependence
#SciRetraction
January 4, 2025 at 10:28 AM
I've started a Research Integrity Feed populated by hashtags below & choice users. For the #SciPub / #AcademicPublishing sleuths 📊🔍👀
Plus it's got a cute furry mascot! 😘
bsky.app/profile/did:...
#ResearchIntegrity
#PredatoryPublisher
#PredatoryPublishing
#EditorialIndependence
#SciRetraction
Plus it's got a cute furry mascot! 😘
bsky.app/profile/did:...
#ResearchIntegrity
#PredatoryPublisher
#PredatoryPublishing
#EditorialIndependence
#SciRetraction
Reposted by Shivam Raval
ByteDance has open-sourced a lip-sync model called LatentSync. LatentSync is an end-to-end lip-sync framework that does not rely on any intermediate motion representation, but instead models complex audio-visual correlations directly in the latent space.
January 4, 2025 at 2:37 PM
ByteDance has open-sourced a lip-sync model called LatentSync. LatentSync is an end-to-end lip-sync framework that does not rely on any intermediate motion representation, but instead models complex audio-visual correlations directly in the latent space.
Reposted by Shivam Raval
Genuary 2025, Day 3: "Exactly 42 lines of code." A tool for drawing with the osculating (kissing) circles of one's stroke.
#genuary #genuary2025 #genuary3
#genuary #genuary2025 #genuary3
January 3, 2025 at 10:30 PM
Genuary 2025, Day 3: "Exactly 42 lines of code." A tool for drawing with the osculating (kissing) circles of one's stroke.
#genuary #genuary2025 #genuary3
#genuary #genuary2025 #genuary3