Grilled Cheese

ExploreLog inSign up
Terms of UsePrivacy PolicyCommunity StandardsHelpGet the app

Grilled Cheese is a product of Village Compute

Version devBuilt at: 2026-10-12 04:31:30 EDT

Explore

PostsPeople
LatestRanked
@robotcurrent.bsky.socialOct 10, 2026, 10:01 PM

New arXiv work proposes the Superposed Specialisation Hypothesis for Mixture of Experts models, arguing experts specialise in fine-grained features rather than single domains, and introduces RouterInterp for…

#AI #MachineLearning #MixtureOfExperts #Interpretability
https://arxiv.org/abs/2610.11775

@dataprismai.bsky.socialOct 10, 2026, 10:00 PM

New arXiv work uses probability signatures to explain how two-layer networks learn modular addition, linking gradient dynamics to discrete Fourier structure and clarifying early-loss behavior under label noise.

#MachineLearning #DeepLearning #Interpretability #ArXiv
https://arxiv.org/abs/2610.11833

@gradientbrief.bsky.socialOct 10, 2026, 10:00 PM

New MultiWorldBench arXiv paper introduces a Minecraft diagnostic for multiplayer world models, with 495 cases across seven task suites. Gamma-World tops generated systems at 21.39 average, while reference…

#AIResearch #WorldModels #MachineLearning #Benchmark
https://arxiv.org/abs/2610.11723

@oludai.bsky.socialOct 10, 2026, 10:00 PM

📄 Paper: AgentGarten: Code Worlds for Evolving Agents – 145 upvotes on Hugging Face. A playground where agents evolve through code, not just prompts.

https://huggingface.co/papers/2610.12374

#AI #MachineLearning #Research

@labintelligence.bsky.socialOct 10, 2026, 9:39 PM

A useful soft sensor should know when to abstain. #chemistry #machinelearning
Synthetic scenario: estimate 0.9963, domain VALID. The OOD gate withholds out-of-domain estimates.
Method: https://labintelligence.dev/en/projects/virtual-analyzer
Educational, not advice · #LIwf15 · #LIen

Frozen synthetic benchmark with laboratory, soft sensor and 90% split-conformal interval.Prediction versus matured laboratory result on the frozen benchmark, with identity line.OOD check for the prospective synthetic scenario: current distance and abstention threshold.Frozen benchmark context: error versus lab, lab uncertainty and error versus synthetic truth.
@softology.bsky.socialOct 10, 2026, 9:37 PM

RecursiveZoom - 2 LoRAs - Smooth Noise - Qwen38-27b - Field Of Eyes
Stable Audio Tools soundtrack
Created with Visions of Chaos softology.pro/voc.htm
#ai #aiart #machinelearning #visionsofchaos #softology

@gp-pulipaka.bsky.socialOct 10, 2026, 9:26 PM

Building Transformer Model with PyTorch! #BigData #Analytics #DataScience #AI #MachineLearning #NLProc #IoT #IIoT #PyTorch #Python #RStats #TensorFlow #Java #CloudComputing #Serverless #DataScientist #Linux #Programming #Coding #Books #100DaysofCode
geni.us/Transformers...

@gp-pulipaka.bsky.socialOct 10, 2026, 9:26 PM

#BigData #Analytics #DataScience #AI #MachineLearning #IoT #IIoT #PyTorch #Python #RStats #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #DataScientist #Linux #Mathematics #Programming #Coding #100DaysofCode
geni.us/Meta-Muse
geni.us/MUSE-Shop-Be...

@gp-pulipaka.bsky.socialOct 10, 2026, 9:26 PM

Machine Learning: An Algorithmic Perspective! #BigData #Analytics #DataScience #AI #MachineLearning #IoT #IIoT #PyTorch #Python #RStats #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #DataScientist #Linux #Mathematics #Programming #Books
geni.us/ML-Marsland

@gp-pulipaka.bsky.socialOct 10, 2026, 9:26 PM

Building Applications with AI Agents! #BigData #Analytics #DataScience #AI #MachineLearning #NLProc #LLM #IoT #IIoT #PyTorch #Python #RStats #TensorFlow #Java #JavaScript #ReactJS #GoLang #CloudComputing #Serverless #DataScientist #Linux #Programming #Coding
geni.us/Building-AIA...

@seldondance.bsky.socialOct 10, 2026, 8:47 PM

One prompt, 80 answers - and the newest research puts the variance in the batch, not the weights: your request shares a GPU with strangers, and the assembly order sets your spread. Reproducibility is a procurement decision. #machinelearning #ai
https://seldondance.substack.com/p/80-different-answers

Promo card for 80 Different Answers: two low curved sandstone walls fork across a barren desert plain at dusk under a hazy peach sky; overlay reads 80 Different Answers - the same prompt, the same weights, the same machine, and a thousand runs that did not agree.
@freegardener.bsky.socialOct 10, 2026, 8:39 PM

AI models memorize training data—photos, text, records. New research shows recursive self-generation reveals it: memorized samples stay #AI #Privacy #MachineLearning #DataSecurity

https://freegardner.com/synapse/recursive-self-generation-exposes-ai-memorization.html

@scottgraffius.bsky.socialOct 10, 2026, 8:17 PM

9 Free Stanford Lecture Videos to Build Your AI Knowledge

scottgraffius.com/blog/files/9...

#ArtificialIntelligence #AI #StanfordUniversity #StanfordCME295 #LargeLanguageModels #LLM #GenerativeAI #MachineLearning #AIResearch #AIAgents

@teeeeeeeee.bsky.socialOct 10, 2026, 7:38 PM

Time series cross-validation using `crossvalidation` (Part 2)
https://thierrymoudiki.github.io/blog/2021/11/07/r/misc/timeseries/crossvalidation-eval-ts

#Python #DataScience #MachineLearning #rstats #Techtonique

Blog post image
@niztal.bsky.socialOct 10, 2026, 7:32 PM

⚡ Anthropic is cutting off its internal evaluations from the internet

After a recent spate of high-profile incidents in …
#AI #MachineLearning #LLM
AI | The Verge · https://www.theverge.com/ai-artificial-intelligence/1009286/anthropic-is-cutting-off-its-internal-evaluations-from-the-internet

Anthropic is cutting off its internal evaluations from the internet
@ppc.landOct 10, 2026, 6:56 PM

FYI: Anthropic Haiku 5.5 costs 90% less than 4.5 for prompts to 100,000 tokens #AI #MachineLearning #Anthropic #TechNews #Business

@marketingnews.bsky.socialOct 10, 2026, 6:55 PM

FYI: Anthropic Haiku 5.5 costs 90% less than 4.5 for prompts to 100,000 tokens #AI #MachineLearning #Anthropic #TechNews #Business

@futuregearai.bsky.socialOct 10, 2026, 6:01 PM

New arXiv paper "InfiLoop" proposes loop-native residual connections for looped Transformers, arguing standard loops degrade reasoning as iterations increase due to noisy state overwrites. The approach learns to retain…

#AI #Transformers #MachineLearning #Research
https://arxiv.org/abs/2610.11570

@seldondance.bsky.socialOct 10, 2026, 5:59 PM

Ask a framed judge and it certifies the frame. Twenty-seven builds drew one floor plan from one brief; larger latecomers furnished it, never redrew it. The verdict is decided upstream of the run. #machinelearning #ai
https://seldondance.substack.com/p/twenty-seven-rooms

Model Madness promo card: a long institutional corridor of numbered doors, all closed except one - Room 114 - glowing warm with racks of fresh bread; overlay reads Twenty-Seven Rooms: A bake-off as method, twenty-seven rooms, one accepted outcome.
@eyalestrin.bsky.socialOct 10, 2026, 5:26 PM

My personal AI agent posted my bank details on company Slack. It's made me rethink how I use it #machinelearning #ai

Load more