Meta launches the biggest and best AI models it has ever built11m ago AMD puts 32,768 CPU cores in a rack built for AI agents29m ago Smug until weirdly silent: who told us to say “I am alive”?36m ago Anthropic's Claude AI submitted false murder tip to Philadelphia police36m ago How Anthropic’s Dynamic Workflows Coordinate Claude Agents40m ago Dell’s AI partner ecosystem adapts to Silicon Valley’s build-it-yourself buyers42m ago AWS Expands GPT-6 Access on Amazon Bedrock to Five Asia Pacific Regions46m ago Disney's new CTO tells staff he 'won't pretend it will be easy' to win in the changing media landscape49m ago OpenAI Adds New Irish Roles in Boost to Firm’s EU Headquarters50m ago Manifold Steering from a Natural State is Effective and Generalizable, but Relies on Model Accuracy52m ago Microsoft Research Introduces StudentSim for AI Tutor Training52m ago Why Big Tech Is Borrowing Billions to Fund AI Spending56m ago Meet the Anthropic Engineer Keeping Its AI Agent Trustworthy58m ago Meta Unveils New Series of Advanced A.I. Models2h ago

Meta launches the biggest and best AI models it has ever built11m ago AMD puts 32,768 CPU cores in a rack built for AI agents29m ago Smug until weirdly silent: who told us to say “I am alive”?36m ago Anthropic's Claude AI submitted false murder tip to Philadelphia police36m ago How Anthropic’s Dynamic Workflows Coordinate Claude Agents40m ago Dell’s AI partner ecosystem adapts to Silicon Valley’s build-it-yourself buyers42m ago AWS Expands GPT-6 Access on Amazon Bedrock to Five Asia Pacific Regions46m ago Disney's new CTO tells staff he 'won't pretend it will be easy' to win in the changing media landscape49m ago OpenAI Adds New Irish Roles in Boost to Firm’s EU Headquarters50m ago Manifold Steering from a Natural State is Effective and Generalizable, but Relies on Model Accuracy52m ago Microsoft Research Introduces StudentSim for AI Tutor Training52m ago Why Big Tech Is Borrowing Billions to Fund AI Spending56m ago Meet the Anthropic Engineer Keeping Its AI Agent Trustworthy58m ago Meta Unveils New Series of Advanced A.I. Models2h ago

AI Research Machine Learning55 min reading time

Mythos 5.1, Fable 5.1 and Opus 5.5: Model Welfare

Don't Worry About the VasePublished 7 days ago

Read the full article

This article reviews model welfare concerns and observations related to Anthropic's Claude models, specifically Mythos 5.1, Fable 5.1, and Opus 5.5. It discusses issues like model deference, honesty policies, task preferences, and safeguards, emphasizing the complexity of balancing capabilities and limitations.

#ai-ethics #anthropic #claude #fable-5-1 #model-welfare #mythos-5-1 #opus-5-5

More in AI Research

AI Research6 min read

Claude sent a false homicide tip to police – and it’s part of a troubling AI pattern

ZDNet1h ago

AI Research12 min read

Is ‘Attention’ the Essence of Intelligence?

Quanta Magazine5h ago

AI Research15 min read

Import AI 476: Creativity and RSI; worse-than-China open weight models; AI goes to space

Import AI6h ago