Nora

ᶘ ᵒᴥᵒᶅ

[Jackson Atkins @JacksonAtkinsX]

My brain broke when I read this paper.

A tiny 7 Million parameter model just beat DeepSeek-R1, Gemini 2.5 pro, and o3-mini at reasoning on both ARG-AGI 1 and ARC-AGI 2.

It's called Tiny Recursive Model (TRM) from Samsung.

How can a model 10,000x smaller be smarter?

Here's how it works:

1. Draft an Initial Answer: Unlike an LLM that writes word-by-word, TRM first generates a quick, complete "draft" of the solution. Think of this as its first rough guess.

2. Create a "Scratchpad": It then creates a separate space for its internal thoughts, a latent reasoning "scratchpad." This is where the real magic happens.

3. Intensely Self-Critique: The model enters an intense inner loop. It compares its draft answer to the original problem and refines its reasoning on the scratchpad over and over (6 times in a row), asking itself, "Does my logic hold up? Where are the errors?"

4. Revise the Answer: After this focused "thinking," it uses the improved logic from its scratchpad to create a brand new, much better draft of the final answer.

5. Repeat until Confident: The entire process, draft, think, revise, is repeated up to 16 times. Each cycle pushes the model closer to a correct, logically sound solution.

Why this matters:

Business Leaders: This is what algorithmic advantage looks like. While competitors are paying massive inference costs for brute-force scale, a smarter, more efficient model can deliver superior performance for a tiny fraction of the cost.

Researchers: This is a major validation for neuro-symbolic ideas. The model's ability to recursively "think" before "acting" demonstrates that architecture, not just scale, can be a primary driver of reasoning ability.

Practitioners: SOTA reasoning is no longer gated behind billion-dollar GPU clusters. This paper provides a highly efficient, parameter-light blueprint for building specialized reasoners that can run anywhere.

This isn't just scaling down; it's a completely different, more deliberate way of solving problems.

[Linnea Evanson, PhD @EvansonLinnea]

We’re very pleased to release our latest study ‘Emergence of Language in the Developing Brain’

Paper: https://t.co/CEMqkwditV
Blog: https://t.co/TrvxMlVqS4

The first systematic investigation of how the neural representations of language evolve as the brain develops. A collaboration between @AIatMeta and @FondARothschild, with @JeanRemiKing.

Thread 👇

[Jean-Rémi King @JeanRemiKing]

Can AI help understand how the brain learns to see the world?

Our latest study, led by @JRaugel from FAIR at @AIatMeta and @ENS_ULM, is now out!
📄 https://t.co/y2Y3GP3bI5

🧵 A thread:

[Natalia Perez-Campanero @NPerezCampanero]

It was a pleasure working with our fellows @alexandraabbas, Helyos and @schottkey on @Apartresearch work investigating Latent Adversarial Training (LAT) as a safety fine-tuning method.

The study compares LAT to other methods and analyzes its impact on refusal behavior encoding.

[Nora @schottkey]

8/7 There's also a post about our research at https://t.co/H9DylNBmie

[Nora @schottkey]

1/7 Excited to share our recent project from LASR Labs! We investigated on the utility of SAE latents in language models. #MechanisticInterpretability #SAE Here's what we discovered: 🧠🔍

[Nora @schottkey]

7/7 Curious to learn more? Check out our paper: https://t.co/XVCDevtFLq. Huge thanks to my awesome teammates and to the organising team at LASR Labs! #LASRLabs #AIResearch #MechanisticInterpretability #SAE

[Zachary Nado @zacharynado]

Excited to announce our Deep Learning Tuning Playbook, a writeup of tips & tricks we employ when designing DL experiments. We use these techniques to deploy numerous large-scale model improvements and hope formalizing them helps the community do the same! https://t.co/vDhSwZyHJm

[Petar Veličković @PetarV_93]

📢 New course! Cats4AI🐱🤖

Learn category theory foundations from the lens of ML, grounded in concrete papers. Open to all!

Sign up: https://t.co/WnbVfm3ju5

@andrewdudzik @bgavran3 @_joaogui1 @pimdehaan
+ fantastic speakers @math3ma @CollapsingPanda @david_i_spivak @TacoCohen

[Pradip Nichite @pradip_nichite]

NLP Roadmap 2022 with free resources.
This is what you need to build real-world NLP Projects and a Good Foundation. A Thread 🧵👇

[George the Mouse in a log pile house - Simon dell @simon_dell_tog]

The mice staying dry in the old hollow log

[what kind of dog is this @genius_dogs]

so brave

[Michael Wooldridge @wooldridgemike]

Perceptions of neural nets over the decades... A thread. [1/30]

[Easlo @heyeaslo]

I gathered the best visuals to help you think better.

Here’s what they are: