·
DataBubble
  • Home
  • Models
  • News
  • Compare
  • Boards
  • Pricing
  • About
  • Newsletter
  • Methodology
  • Contact
Latest
Meta’s Adam Mosseri says AI token budgets could soon be capped per engineer1h◆Google Images gets a Pinterest-like redesign focused on discovery1h◆The Google Images homepage will recommend photos even before you search1h◆Celebrating 25 years of visual search innovation1h◆Spotify is now an AI chatbot, too1h◆New York State halts construction of all new data centers2h◆Reflection inks $1B compute deal with Nebius3h◆The real AI race may no longer be at the frontier3h◆Spotify expands its AI push with a ChatGPT-like music assistant3h◆Sam Altman didn’t need another lawsuit3h◆Superhuman’s new auto-draft feature almost makes me like AI replies3h◆Google’s Demis Hassabis says it’s time for a global AI watchdog — led by the US5h◆New York becomes the first state to enact a data center moratorium8h◆FAD-SA-GRU: Enhancing Hate Speech Detection in Algerian Dialect Through Feature-Augmented Self-Attention GRU Networks13h◆SMETA-ZSL:Semantic Meta-Alignment for Zero-Shot Threat Classification13h◆Dzongkha Next Word Prediction System13h◆MoP-JEPA: Hard-Assigned Predictor Mixtures for Stochastic JEPA World Models13h◆Graph Neural Networks for RFID-Based Spatial Geometry Inference in Spatial AI Systems13h◆SpurLens: Automatic Detection of Spurious Cues in Multimodal LLMs13h◆Constraint-Aware Hierarchical Search for Regulation-Driven Fine-Grained Classification13h◆Meta’s Adam Mosseri says AI token budgets could soon be capped per engineer1h◆Google Images gets a Pinterest-like redesign focused on discovery1h◆The Google Images homepage will recommend photos even before you search1h◆Celebrating 25 years of visual search innovation1h◆Spotify is now an AI chatbot, too1h◆New York State halts construction of all new data centers2h◆Reflection inks $1B compute deal with Nebius3h◆The real AI race may no longer be at the frontier3h◆Spotify expands its AI push with a ChatGPT-like music assistant3h◆Sam Altman didn’t need another lawsuit3h◆Superhuman’s new auto-draft feature almost makes me like AI replies3h◆Google’s Demis Hassabis says it’s time for a global AI watchdog — led by the US5h◆New York becomes the first state to enact a data center moratorium8h◆FAD-SA-GRU: Enhancing Hate Speech Detection in Algerian Dialect Through Feature-Augmented Self-Attention GRU Networks13h◆SMETA-ZSL:Semantic Meta-Alignment for Zero-Shot Threat Classification13h◆Dzongkha Next Word Prediction System13h◆MoP-JEPA: Hard-Assigned Predictor Mixtures for Stochastic JEPA World Models13h◆Graph Neural Networks for RFID-Based Spatial Geometry Inference in Spatial AI Systems13h◆SpurLens: Automatic Detection of Spurious Cues in Multimodal LLMs13h◆Constraint-Aware Hierarchical Search for Regulation-Driven Fine-Grained Classification13h◆
News/What Does the Weight Norm Control in Grokking? Logit-Scale Mediation under Cross-Entropy
arxiv
PublishedJune 18, 2026 at 4:00 AM
—neutral

What Does the Weight Norm Control in Grokking? Logit-Scale Mediation under Cross-Entropy

Source
arxiv.orgfull article ↗
Read on arxiv→
Publisher summary· verbatim

arXiv:2606.18465v1 Announce Type: cross Abstract: Grokking, the delayed jump from memorization to generalization, is usually tied to the weight norm: a smaller norm generalizes sooner. We ask what the norm actually controls. Holding the weight norm fixed by clamping and varying only an output temper

Stay posted· Newsletter

A 5-min weekly brief — top movers, price watch, story of the week.

// no spam · unsubscribe one-click · free forever

Discussion
Source
↗
arxiv
Read original ↗All from arxiv →

No replies yet. Be first.

Source
↗
arxiv
Read original ↗All from arxiv →

Related coverage

More from ARXIV
arxivFAD-SA-GRU: Enhancing Hate Speech Detection in Algerian Dialect Through Feature-Augmented Self-Attention GRU Networks13harxivSMETA-ZSL:Semantic Meta-Alignment for Zero-Shot Threat Classification13harxivDzongkha Next Word Prediction System13harxivMoP-JEPA: Hard-Assigned Predictor Mixtures for Stochastic JEPA World Models13h
The Bubble Brief
WEEKLY

Read AI insights every Tuesday — top movers, new releases, story of the week.

// no spam · unsubscribe one-click · free forever

Originally published on arxiv ↗
HomeModelsNews