Gemma-4-E4B

mentions 2 type Organization feed RSS

// recent coverage 2 mentions

04:00

2026-09-10

arxiv.org

large-language-models

The Failure Happens Before the Drift: The Social Dynamics of Values in LLM Agent Societies

A World Values Survey-grounded simulation of LLM agent societies found that more than 50% of personas fail to express their assigned WVS value profiles from the outset, while only 2-7% drift after rep…

04:00

2026-08-11

machinebrief.com

artificial-intelligence

Do All LLMs Know When They're Being Harmful? A Reproducibility Study of Latent-Space Safety Probes Across Model Families

A reproducibility study of latent-space safety probes found that lightweight MLP probes trained on final-layer activations of LLaMA-3.1-8B detect harmful prompts with F1 scores within 0.37 percentage …

// co-occurs with top 8 entities

arXiv 2 LLaMA-3.1-8B 1 Mistral-7B-v0.3 1 Qwen2-7B 1 WildJailbreak 1 BeaverTails 1 AEGIS 2.0 1 World Values Survey 1

// topics top 6 topics

ai safety 2 ai research 2 artificial intelligence 1 machine learning 1 large language models 1 ai agents 1