iiinova — where your digital self keeps evolving

Reposted by Joel Chen

Ai2 @ai2.bsky.social · 16/09/2025

🚀 Introducing Fluid Benchmarking—an adaptive way to evaluate LLMs. Inspired by psychometrics, it tailors which questions to ask based on each model’s capability, making evals more efficient & reliable. 🧵

1207

Reposted by Joel Chen

Ai2 @ai2.bsky.social · 31/07/2025

New interview drop! ⚠️ Ai2 Senior Research Scientist @natolambert.bsky.social sat down with @kylelwiggers.bsky.social, Ai2’s new Comms Lead, to talk about what sets Ai2 apart from other frontier AI labs, and our role in the research community. Watch the highlights. 👇

182

Reposted by Joel Chen

Sung Kim @sungkim.bsky.social · 01/08/2025

OpenAI's open-weight model may be a MoE with 120B parameters with 20B active.

2172

Reposted by Joel Chen

Samuel Vaiter @samuelvaiter.com · 13/12/2024

Automatic differentiation in forward mode computes derivatives by breaking down functions into elem operations and propagating derivatives alongside values. It’s efficient for functions with fewer inputs than outputs and for Jacobian-vect prod, using for instance dual numbers.

23810

Reposted by Joel Chen

Thomas Wolf @thomwolf.bsky.social · 11/12/2024

The Open LLM Leaderboard got a new front page for Christmas Check it out at huggingface.co/spaces/open-...

26612

Reposted by Joel Chen

Thomas Wolf @thomwolf.bsky.social · 08/12/2024

The FineWeb team is happy to finally release "FineWeb2" 🥂🥳 FineWeb 2 extends the data driven approach to pre-training dataset design that was introduced in FineWeb 1 to now covers 1893 languages/scripts Details: huggingface.co/datasets/Hug... A detailed open-science tech report is coming soon

310613

Reposted by Joel Chen

Nathan Lambert @natolambert.bsky.social · 05/12/2024

The title of our NeurIPS Tutorial is "Opening the Language Model Pipeline: A Tutorial on Data Preparation, Model Training, and Adaptation" Compared to a 2023 RLHF tutorial I gave at ICML we know SO MUCH more about all steps of the training process. 930am PST, West Ballroom B. Let's go!!

1443

Reposted by Joel Chen

Yoav Artzi @yoavartzi.com · 02/12/2024

I am seriously behind uploading Learning Machines videos, but I did want to get @jonathanberant.bsky.social's out sooner than later. It's not only a great talk, it also gives a remarkably broad overview and contextualization, so it's an excellent way to ramp up on post-training youtu.be/2AthqCX3h8U

youtu.be

Jonathan Berant (Tel Aviv University / Google) / Towards Robust Language Model Post-training

YouTube video by Yoav Artzi

15312

Reposted by Joel Chen

Nathan Lambert @natolambert.bsky.social · 01/12/2024

To be clear this is a filter running on top of the base LLM and not some weird training. Very similar to safety and other similar things.

4251

Reposted by Joel Chen

Juan Diego Rodriguez @juand-r.bsky.social · 01/12/2024

There's too many starter packs. 👇 Here's a list, mostly for NLP, ML, and related areas.

go.bsky.app

NLP grad students

Join the conversation

34211

Reposted by Joel Chen

Lucas Beyer (bl16) @giffmana.ai · 29/11/2024

Some recent discussions made me write up a short read on how I think about doing computer vision research when there's clear potential for abuse. Alternative title: why I decided to stop working on tracking. Curious about other's thoughts on this. lb.eyer.be/s/cv-ethics....

1917320

Reposted by Joel Chen

Yoav Goldberg @yoavgo.bsky.social · 29/11/2024

are there works (in RL? in planning?) that characterize different long-horizon tasks in terms of their hardness? for example, how reactive can you be? how often if at all do you need to backtrack or change strategies? how delayed is the feedback? etc. is there a taxonomy or theory of these?

6608

Reposted by Joel Chen

Jia-Bin Huang @jbhuang0604.bsky.social · 01/12/2024

3D Gaussian Splatting is awesome! BUT, the 3D Gaussian primitives often have difficulty reconstructing fine appearance details. 😩 Meet Textured Gaussians! ✋ Our RGBA Textured Gaussians enhance 3D scene appearance modeling. Learn more at textured-gaussians.github.io

58214

Reposted by Joel Chen

Yoav Goldberg @yoavgo.bsky.social · 01/12/2024

what did y'all imagine machine learning is, before learning it and being indoctrinated that it all about optimizing parameterized functions involving many matrices?

23262

Reposted by Joel Chen

Nathan Lambert @natolambert.bsky.social · 29/11/2024

The most realistic reason to be pro open source AI is to reduce concentration of power.

619036

Reposted by Joel Chen

Chris Offner @chrisoffner3d.bsky.social · 28/11/2024

Welcome to all the newcomers! Here's a starter pack for the 3D Computer Vision community: go.bsky.app/Cfm9XFe

74612

Reposted by Joel Chen

Chris Offner @chrisoffner3d.bsky.social · 27/11/2024

The iPhone LiDAR depth is pretty stable but low-res and low detail. Monodepth is highly detailed but unstable ("depth flickering"). For more stable and detailed metric depth, I solve for the per-frame affine transform that optimally "anchors" the monodepth to the LiDAR. youtube.com/shorts/u3OVj...

youtube.com

Anchoring monodepth to LiDAR depth | Pedestrian walk

YouTube video by Chris Offner

3594

Reposted by Joel Chen

Nathan Sarrazin @nsarrazin.com · 28/11/2024

We just deployed Qwen/QwQ-32B-Preview on HuggingChat! It's Qwen's latest experimental reasoning model. It's super interesting to see the reasoning steps, and with really impressive results too. Feel free to try it out here: huggingface.co/chat/models/... I'd love to get your feedback on it!

huggingface.co

Qwen/QwQ-32B-Preview - HuggingChat

Use Qwen/QwQ-32B-Preview with HuggingChat

1397

Reposted by Joel Chen

Juliet Shen @julietshen.online · 28/11/2024

Just FYI because it seems relevant and I've seen it misstated a few times, LAION retrained their dataset and provided a diff to migrate over to the filtered Re-LAION-5B dataset HuggingFace, like most platforms that handle user content, do checks for CSAM too. laion.ai/blog/relaion...

laion.ai

Releasing Re-LAION 5B: transparent iteration on LAION-5B with additional safety fixes | LAION

<p>Today, following <a href="https://laion.ai/notes/laion-maintenance/">a safety revision procedure</a>, we announce Re-LAION-5B, an updated version of LAION...

1125

Reposted by Joel Chen

Noa Garcia @noagarciad.bsky.social · 22/11/2024

As I haven't found it out there yet, I made the Women in computer vision started pack. Many more missing, please let me know how is already in bsky to add them! go.bsky.app/BowzivT

114314

Reposted by Joel Chen

🤖ScottsRobots🤖 @scottsrobots.bsky.social · 14/11/2024

If you like robots & makers here's a thread starter packs! go.bsky.app/9xL642E

93211

Reposted by Joel Chen

Philipp Schmid @philschmid.bsky.social · 25/11/2024

Created a visual for how function calling works. Wdyt? 🤔

6242

Reposted by Joel Chen

Kosta Derpanis @csprofkgd.bsky.social · 24/11/2024

Here we grow again. Introducing my 2nd AWESOME #ComputerVision Starter Pack. Missed you? Let me know in the thread. go.bsky.app/Gjcs339

457919

Reposted by Joel Chen

vb @reach-vb.hf.co · 24/11/2024

Massive week for Open Source AI/ ML Mistral Pixtral & Instruct Large - ~123B, 128K context, multilingual, json + function calling & open weights Allen AI Tülu 70B & 8B - competive with claude 3.5 haiku, beats all major open models like llama 3.1 70B, qwen 2.5 and nemotron

3559

Reposted by Joel Chen

Maria Antoniak @mariaa.bsky.social · 24/11/2024

I kinda think this #NLP starter pack has served and now outlived its original purpose, but I'm not sure what to do with it. Delete? Rename? What do you all think? Also, if anyone wants to be removed now, just let me know! go.bsky.app/SngwGeS

9385

Reposted by Joel Chen

Daniel Lowd @dlowd.com · 24/11/2024

This is important work — analyzing how automated systems influence lives on a massive scale. Every system needs a practical way to appeal to a human with power to override the system. I do object to using the term “AI” — this report targets ALL automated systems, including if/else statements.

281

Reposted by Joel Chen

fizz ☭ @fizz.allura.moe · 24/11/2024

open source tulu 3 model recreation! rivals the original sft and other models in its size range huggingface.co/allura-org/T...

huggingface.co

allura-org/Teleut-7b · Hugging Face

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

1346

Reposted by Joel Chen

Nacho Sanguinetti @neuroetho.bsky.social · 24/11/2024

Just In case people need to follow people in neuroscience focused on natural behaviors and the brains of cool non-trad animals! Please join let me know if you wanna be added. go.bsky.app/AcLpM4h

4011053

Reposted by Joel Chen

Quanquan Gu @quanquangu.bsky.social · 22/11/2024

In addition to the Deep Learning Theory starter pack, I've also put together a starter pack for Reinforcement Learning Theory. Let me know if you'd like to be included or suggest someone to add to the list! go.bsky.app/LWyGAAu

112910

Reposted by Joel Chen

Thomas Wolf @thomwolf.bsky.social · 22/11/2024

Excited to release SmolTalk: the secret recipe behind the best-in-class performance of SmolLM2 hf.co/datasets/Hug... The SmolLM team at Hugging Face curated new synthetic datasets to enhance instruction following and reasoning. SmolTalk also supports rewriting, summarization, and function calling.

61324

Reposted by Joel Chen

Sung Kim @sungkim.bsky.social · 21/11/2024

Free eBook: Machine Learning Systems by Vijay Janapa Reddi Principles and Practices of Engineering Artificially Intelligent Systems mlsysbook.ai

1304

Reposted by Joel Chen

Chris Offner @chrisoffner3d.bsky.social · 21/11/2024

After my general computer vision starter pack is now full (150/150 entries reached), here is one specific to 3D Vision: go.bsky.app/Cfm9XFe

1010529

Reposted by Joel Chen

Carlos E. Perez (IntuitMachine) @ceperez.bsky.social · 21/11/2024

DeepSeek-Prover Outperforms GPT-4 by Using Agentic AI to Generate and Validate Synthetic Data arxiv.org/abs/2408.08152

0124

Reposted by Joel Chen

Nathan Lambert @natolambert.bsky.social · 20/11/2024

Very impressive plot from DeepSeek AI - DeepSeek-R1-Lite-Preview. One step closer to a replication. I need to be able to see these weird RL data recipes, then I will give my open replication of o1 stamp of approval. Being potentially soon! Just demo for now buff.ly/4hUTWFc

1473

Reposted by Joel Chen

Yoshitomo Matsubara @yoshitomo-matsubara.net · 21/11/2024

Here is a list of ML OSS & Open Source / Science enthusiasts I found on Bluesky 🦋 go.bsky.app/8MFcfXd Let me know if you find such people here! I'm still new here and probably the list misses many must-add people, so let's built it together💪

4011249

Reposted by Joel Chen

Leandro von Werra @lvwerra.bsky.social · 19/11/2024

All the things you need to know to pretrain an LLM at home*! Gave a workshop at Uni Bern: starts with scaling laws and goes to web scale data processing and finishes training with 4D parallelism and ZeRO. *assuming your home includes an H100 cluster

5779

Reposted by Joel Chen

Sung Kim @sungkim.bsky.social · 20/11/2024

Chai-1r, an updated version of Chai-1 with full support for restraints. Prompting Chai-1r with prior structural knowledge can boost accuracy by >2x.

1131

Reposted by Joel Chen

Kosta Derpanis @csprofkgd.bsky.social · 19/11/2024

My growing list of #computervision researchers on Bsky. Missed you? Let me know. go.bsky.app/M7HGC3Y

8813242

Reposted by Joel Chen

Grzegorz Chrupała @grzegorz.chrupala.me · 19/11/2024

I've started putting together a starter pack with people working on Speech Technology and Speech Science: go.bsky.app/BQ7mbkA (Self-)nominations welcome!

448234

Reposted by Joel Chen

Chris Offner @chrisoffner3d.bsky.social · 17/11/2024

Welcome to all new arrivals here on Bluesky! :) Here's a starter pack of people working on computer vision. go.bsky.app/PkAKJu5

219733

Reposted by Joel Chen

Katherine Liu @katherineliu.bsky.social · 19/11/2024

A starter pack of starter packs: Robotics and AI go.bsky.app/DfAoaJ1 Computer Vision go.bsky.app/PkAKJu5 Computer Graphics Research go.bsky.app/ckQ1u9 Grumpy Machine Learners go.bsky.app/6ddpivr Reinforcement Learning go.bsky.app/3WPHcHg

79628

Reposted by Joel Chen

Jakub M. Tomczak @jmtomczak.bsky.social · 19/11/2024

I've created a startepack on Generative Modeling: go.bsky.app/Hd9ykTw

178626

Reposted by Joel Chen

lebellig @lebellig.bsky.social · 18/11/2024

I created 3 introductory notebooks on Flow Matching models to help get started with this exciting topic! ✨ 1. Annotated Flow Matching paper: github.com/gle-bellier/... 2. Discrete Flow Matching: github.com/gle-bellier/... 3. Minimal FM in Jax: github.com/gle-bellier/...

github.com

GitHub - gle-bellier/flow-matching: Annotated Flow Matching paper

Annotated Flow Matching paper. Contribute to gle-bellier/flow-matching development by creating an account on GitHub.

314821