iiinova — where your digital self keeps evolving

Reposted by Koustuv Sinha

Adina Williams @adinawilliams.bsky.social · 15/07/2025

Our team is hiring a postdoc in (mechanistic) interpretability! The ideal candidate will have research experience in interpretability for text and/or image generation models and be excited about open science! Please consider applying or sharing with colleagues: metacareers.com/jobs/2223953961352324

careers.com

0115

Reposted by Koustuv Sinha

Benno Krojer @bennokrojer.bsky.social · 13/06/2025

Excited to share the results of my recent internship! We ask 🤔 What subtle shortcuts are VideoLLMs taking on spatio-temporal questions? And how can we instead curate shortcut-robust examples at a large-scale? We release: MVPBench Details 👇🔬

1165

Reposted by Koustuv Sinha

Alexander Doria @dorialexander.bsky.social · 19/02/2025

The HuggingFace/Nanotron team just shipped an entire pretraining textbook in interactive format. huggingface.co/spaces/nanot... It’s not just a great pedagogic support, but many unprecedented data and experiments presented for the first time in a systematic way.

0409

Reposted by Koustuv Sinha

Kaitlyn Zhou @kaitlynzhou.bsky.social · 19/02/2025

Excited to have two papers at #NAACL2025! The first reveals how human over-reliance can be exacerbated by LLM friendliness. The second presents a novel computational method for concept tracing. Check them out! arxiv.org/pdf/2407.07950 arxiv.org/pdf/2502.05704

2276

Koustuv Sinha @koustuvsinha.com · 20/02/2025

Congrats, nice and refreshing papers, especially the word confusion idea! We need better similarity methods, good to see developments in this front! Curious if the confusion similarity depends on the label size of the classifier?

000

Reposted by Koustuv Sinha

fairseq2 @fairseq2.bsky.social · 12/02/2025

👋 Hello world! We’re thrilled to announce the v0.4 release of fairseq2 — an open-source library from FAIR powering many projects at Meta. pip install fairseq2 and explore our trainer API, instruction & preference finetuning (up to 70B), and native vLLM integration.

142

Koustuv Sinha @koustuvsinha.com · 11/02/2025

Many many congratulations!! 🥳🎉🎉

120

Koustuv Sinha @koustuvsinha.com · 02/02/2025

another factor which makes simple mlps work is visual token length. if you care about shorter tokens, you need a better mapper. these days most llms are capable of long context, which reduces the need of compressing visual tokens.

130

Koustuv Sinha @koustuvsinha.com · 02/02/2025

one hypothesis why simple mappers work is 1. unfreezing the LLM provides enough parameters for mapping, 2. richer vision representations are closer to llm internal latent space arxiv.org/abs/2405.07987

120

Koustuv Sinha @koustuvsinha.com · 02/02/2025

good questions! from what I see some folks still use complex mappers like Perceivers, but often simple mlp works good enough. the variable which induces the biggest improvement is almost always the alignment data.

110

Koustuv Sinha @koustuvsinha.com · 31/01/2025

This is actually a cool result - token length being a rough heuristic for confidence of models?

010

Reposted by Koustuv Sinha

Anna Rogers @annarogers.bsky.social · 14/01/2025

I am shocked by the death of Felix Hill. He was one of the brightest minds of my generation. His last blog post on the stress of working in AI is very poignant. Apart from the emptiness of working mostly to make billionaires even richer, there's the intellectual emptiness of 'scale is all you need'

0398

Koustuv Sinha @koustuvsinha.com · 26/12/2024

Lots of cool findings in our paper as well as in the website: tsb0601.github.io/metamorph/ Excited to see how the community "MetaMorph"'s existing LLMs!

040

Koustuv Sinha @koustuvsinha.com · 26/12/2024

We posted our paper on arxiv recently, sharing this here too: arxiv.org/abs/2412.141... - work led by our amazing intern Peter Tong. Key findings: - LLMs can be trained to generate visual embeddings!! - VQA data appears to help a lot in generation! - Better understanding = better generation!

tsb0601.github.io

MetaMorph: Multimodal Understanding and Generation via Instruction Tuning

180

Koustuv Sinha @koustuvsinha.com · 17/12/2024

I wonder if veo-2 would be better at these prompts!

230

Koustuv Sinha @koustuvsinha.com · 13/12/2024

Co-organized by @randomwalker.bsky.social @peterhenderson.bsky.social, @in4dmatics.bsky.social Naila Murray, @adinawilliams.bsky.social, Angela Fan, Mike Rabbat and Joelle Pineau. Checkout our website for CFP and more details: reproml.org

reproml.org

MLRC 2025

Machine Learning Reproducibility Challenge

010

Koustuv Sinha @koustuvsinha.com · 13/12/2024

🚨 We are pleased to announce the first, in-person event for the Machine Learning Reproducibility Challenge, MLRC 2025! Save your dates: August 21st, 2025 at Princeton!

3101

Reposted by Koustuv Sinha

Adina Williams @adinawilliams.bsky.social · 11/12/2024

Our paper PRISM alignment won a best paper award at #neurips2024! All credits to @hannahrosekirk.bsky.social A.Whitefield, P.Röttger, A.M.Bean, K.Margatina, R.Mosquera-Gomez, J.Ciro, @maxbartolo.bsky.social H.He, B.Vidgen, S.Hale Catch Hannah tomorrow at neurips.cc/virtual/2024/poster/97804

blog.neurips

2679

Koustuv Sinha @koustuvsinha.com · 10/12/2024

Also, MLRC is now in 🦋 as well - do follow! :) @reproml.org

000

Koustuv Sinha @koustuvsinha.com · 10/12/2024

Checkout the MLRC 2023 posters at #NeurIPS 2024 this week: reproml.org/proceedings/ - do drop by to these posters and say hi!

reproml.org

Online Proceedings | MLRC

Machine Learning Reproducibility Challenge

100

Reposted by Koustuv Sinha

Andrei Bursuc @abursuc.bsky.social · 22/11/2024

The return of the Autoregressive Image Model: AIMv2 now going multimodal. Excellent work by @alaaelnouby.bsky.social & team with code and checkpoints already up: arxiv.org/abs/2411.14402

1468

Koustuv Sinha @koustuvsinha.com · 21/11/2024

Yes, that imo is one of the most exciting outcome for this direction - learning a new modality with much less compute. We have some really nice results, can’t wait to share it with everyone, stay tuned!

110

Reposted by Koustuv Sinha

Michael J. Black @michael-j-black.bsky.social · 20/11/2024

For those who missed this post on the-network-that-is-not-to-be-named, I made public my "secrets" for writing a good CVPR paper (or any scientific paper). I've compiled these tips of many years. It's long but hopefully it helps people write better papers. perceiving-systems.blog/en/post/writ...

perceiving-systems.blog

Writing a good scientific paper

426065

Reposted by Koustuv Sinha

Laura @lauraruis.bsky.social · 20/11/2024

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 ⚙️🔢 🧵⬇️

36850139

Koustuv Sinha @koustuvsinha.com · 20/11/2024

When I first read this paper, I instinctively scoffed at the idea. But the more I look at empirical results, the more I’m convinced this paper highlights something fundamentally amazing. Lots of exciting research on this direction will come very soon! arxiv.org/abs/2405.07987

330

Reposted by Koustuv Sinha

Itai Yanai @itaiyanai.bsky.social · 09/11/2024

Doing good science is 90% finding a science buddy to constantly talk to about the project.

22877215

Koustuv Sinha @koustuvsinha.com · 17/11/2024

Same here! Lets make a club! 😅

000

Reposted by Koustuv Sinha

Maike Osborne @maosbot.bsky.social · 09/11/2024

New here? Interested in AI/ML? Check out these great starter packs! AI: go.bsky.app/SipA7it RL: go.bsky.app/3WPHcHg Women in AI: go.bsky.app/LaGDpqg NLP: go.bsky.app/SngwGeS AI and news: go.bsky.app/5sFqVNS You can also search all starter packs here: blueskydirectory.com/starter-pack...

66558212