AI models on realistic cyber ranges
…We have continued collaborating with Incalmo to run evaluations on these cyber ranges (simulated network environments for security testing). A notable development during the testing of Claude Sonnet 4.5 is that…
Search
Related topics:
Claude
People also ask
Why does agentic misalignment happen?
Before we started this research, it was not clear where the misaligned behavior was coming from. Our main two hypotheses were: Our post-training process was accidentally encouraging this behavior with misaligned rewards.This behavior was coming from the pre-trained model and our post-training was failing to sufficiently discourage it. We now believe that (2) is largely responsible. Specifically, at the time of Claude 4’s training, the vast majority of our alignment training was standard chat-based Reinforcement Learning from Human Feedback RLHF data that did not include any agentic tool use. T
Teaching Claude whyTop stories
An update on our election safeguards
…Agents for financial services We're releasing ten new Cowork and Claude Code plugins, integrations with the Microsoft 365 suite, new connectors, and an MCP app for financial services and insurance organizations…
Anthropic’s Long-Term Benefit Trust appoints Vas Narasimhan to Board of Directors
…Agents for financial services We're releasing ten new Cowork and Claude Code plugins, integrations with the Microsoft 365 suite, new connectors, and an MCP app for financial services and insurance organizations…
AI agents find smart contract exploits
We evaluated AI agents' ability to exploit smart contracts using a new benchmark comprising contracts that were actually exploited. On contracts exploited after the latest knowledge cutoffs, Claude Opus 4.5, Claude…
Developing Nuclear Safeguards for AI
…We have already deployed this classifier on Claude traffic as part of our broader system for identifying misuse of our models. Early deployment data suggests the classifier works well with real Claude…
A global workspace in language models
…the capital, the language, the continent, and the currency. Then we swapped “France” for “China” in the J-space, with the exact same intervention in each context. Claude answered with “Beijing,” “Chinese…
Progress from our Frontier Red Team
…This improvement in cyber capabilities has continued with our latest model, Claude 3.7 Sonnet. On Cybench — a public benchmark using CTF challenges to evaluate LLMs — Claude 3.7 Sonnet solves about…
Focus areas for The Anthropic Institute
…externalities? Related content Teaching Claude why New research on how we've reduced agentic misalignment. Natural Language Autoencoders: Turning Claude’s thoughts into text AI models like Claude talk in words but…
Natural Language Autoencoders
…In practice, we've already used NLAs in the pre-deployment alignment audits of Claude Mythos Preview and Claude Opus 4.6 . We are excited to continue using NLAs in our investigations…
Making Claude a chemist
…Looking ahead As we continue to improve Claude’s performance in chemistry, we are focusing specifically on a handful of bottlenecks that slow chemists down the most. Reading and rendering chemical structures…